Skip to content
Changelog

Report a bug

Metabase integration: operations for AI agents

Metabase connects to Type.com as an integration. This page lists the 699 operations documented for Claude and Codex in Type.com: 339 read, 250 write, 110 destructive.

About Metabase

Metabase vendor API references, subject to vendor permissions and account availability.

Categories: Analytics & Data

How to connect Metabase to Type.com

Access
Organization connection: one shared account for the workspace.

To add the connection, follow the steps in Space connections.

Documented operations

699 operations. Your connection's permissions determine which operations are available. API references describe the vendor surface; they do not expand your connection's credentials or access.

699 operations listed.

  • Read

    GET /api/action

    Returns actions that can be used for QueryActions. By default lists all viewable actions. Pass optional `?model-id=<model-id>` to limit to actions on a particular model.

    GET /api/action

  • Read

    GET /api/action/public

    Fetch a list of Actions with public UUIDs. These actions are publicly-accessible *if* public sharing is enabled.

    GET /api/action/public

  • Read

    GET /api/action/{action-id}

    Fetch an Action.

    GET /api/action/{action-id}

  • Read

    GET /api/activity/most_recently_viewed_dashboard

    Get the most recently viewed dashboard for the current user. Returns a 204 if the user has not viewed any dashboards in the last 24 hours.

    GET /api/activity/most_recently_viewed_dashboard

  • Read

    GET /api/activity/recent_views

    Get a list of 100 models (cards, models, tables, dashboards, and collections) that the current user has been viewing most recently. Return a maximum of 20 model of each, if they've looked at at least 20.

    GET /api/activity/recent_views

  • Read

    GET /api/activity/recents

    Get a list of recent items the current user has been viewing most recently under the `:recents` key. Allows for filtering by context: views or selections

    GET /api/activity/recents

  • Read

    GET /api/agent/v1/ping

    Health check endpoint for the Agent API.

    GET /api/agent/v1/ping

  • Read

    GET /api/alert

    Fetch alerts which the current user has created or will receive, or all alerts if the user is an admin. The optional `user_id` will return alerts created by the corresponding user, but is ignored for non-admin users.

    GET /api/alert

  • Read

    GET /api/alert/{id}

    Fetch an alert by ID

    GET /api/alert/{id}

  • Read

    GET /api/analytics/anonymous-stats

    Anonymous usage stats. Endpoint for testing, and eventually exposing this to instance admins to let them see what is being phoned home.

    GET /api/analytics/anonymous-stats

  • Read

    GET /api/api-key

    Get a list of API keys with the default scope. Non-paginated.

    GET /api/api-key

  • Read

    GET /api/api-key/count

    Get the count of API keys in the DB with the default scope.

    GET /api/api-key/count

  • Read

    GET /api/apps

    List the data apps provided by the connected repository. Pass `available=true` to return only enabled apps without sync errors.

    GET /api/apps

  • Read

    GET /api/apps/repo-status

    Status of the connected repository as it relates to data apps. Data apps are pulled by the remote-sync import (manual "Pull changes", auto-import, or startup); the connection is configured on the remote-sync settings page.

    GET /api/apps/repo-status

  • Read

    GET /api/apps/sandbox-host

    Serve the document used as the `src` of the Near-Membrane realm iframe, carrying a per-document CSP that confines `'unsafe-eval'` to that realm. See [[sandbox-host-csp]].

    GET /api/apps/sandbox-host

  • Read

    GET /api/apps/{slug}

    Fetch metadata for a single enabled data app by its slug.

    GET /api/apps/{slug}

  • Read

    GET /api/apps/{slug}/bundle

    Serve the cached JS bundle for a single enabled data app by slug. Honors `If-None-Match` against the content-hash ETag with a 304.

    GET /api/apps/{slug}/bundle

  • Read

    GET /api/automagic-dashboards/database/{id}/candidates

    Return a list of candidates for automagic dashboards ordered by interestingness.

    GET /api/automagic-dashboards/database/{id}/candidates

  • Read

    GET /api/automagic-dashboards/model_index/{model-index-id}/primary_key/{pk-id}

    Return an automagic dashboard for the record of the model indexed by `model-index-id` whose primary key is `pk-id`. The record's value, used in the title, is read through the QP as the requesting user (never from `model_index_value`, which is not permission-checked); a `pk-id` they cannot resolve yields a dashboard titled by the pk, with no matching rows, rather than a 404.

    GET /api/automagic-dashboards/model_index/{model-index-id}/primary_key/{pk-id}

  • Read

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}

    Return an automagic dashboard for entity `entity` with id `id`.

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}

  • Read

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/cell/{cell-query}

    Return an automagic dashboard analyzing cell in automagic dashboard for entity `entity` defined by query `cell-query`.

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/cell/{cell-query}

  • Read

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/cell/{cell-query}/compare/{comparison-entity}/{comparison-entity-id-or-query}

    Return an automagic comparison dashboard for cell in automagic dashboard for entity `entity` with id `id` defined by query `cell-query`; compared with entity `comparison-entity` with id `comparison-entity-id-or-query.`.

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/cell/{cell-query}/compare/{comparison-entity}/{comparison-entity-id-or-query}

  • Read

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/cell/{cell-query}/rule/{prefix}/{dashboard-template}

    Return an automagic dashboard analyzing cell in question with id `id` defined by query `cell-query` using dashboard-template `dashboard-template`.

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/cell/{cell-query}/rule/{prefix}/{dashboard-template}

  • Read

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/cell/{cell-query}/rule/{prefix}/{dashboard-template}/compare/{comparison-entity}/{comparison-entity-id-or-query}

    Return an automagic comparison dashboard for cell in automagic dashboard for entity `entity` with id `id` defined by query `cell-query` using dashboard-template `dashboard-template`; compared with entity `comparison-entity` with id `comparison-entity-id-or-query.`.

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/cell/{cell-query}/rule/{prefix}/{dashboard-template}/compare/{comparison-entity}/{comparison-entity-id-or-query}

  • Read

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/compare/{comparison-entity}/{comparison-entity-id-or-query}

    Return an automagic comparison dashboard for entity `entity` with id `id` compared with entity `comparison-entity` with id `comparison-entity-id-or-query.`

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/compare/{comparison-entity}/{comparison-entity-id-or-query}

  • Read

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/query_metadata

    Return all metadata for an automagic dashboard for entity `entity` with id `id`.

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/query_metadata

  • Read

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/rule/{prefix}/{dashboard-template}

    Return an automagic dashboard for entity `entity` with id `id` using dashboard-template `dashboard-template`.

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/rule/{prefix}/{dashboard-template}

  • Read

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/rule/{prefix}/{dashboard-template}/compare/{comparison-entity}/{comparison-entity-id-or-query}

    Return an automagic comparison dashboard for entity `entity` with id `id` using dashboard-template `dashboard-template`; compared with entity `comparison-entity` with id `comparison-entity-id-or-query.`.

    GET /api/automagic-dashboards/{entity}/{entity-id-or-query}/rule/{prefix}/{dashboard-template}/compare/{comparison-entity}/{comparison-entity-id-or-query}

  • Read

    GET /api/bookmark

    Fetch all bookmarks for the user

    GET /api/bookmark

  • Read

    GET /api/bug-reporting/connection-pool-details

    Returns database connection pool info for the current Metabase instance.

    GET /api/bug-reporting/connection-pool-details

  • Read

    GET /api/bug-reporting/details

    Returns version and system information relevant to filing a bug report against Metabase.

    GET /api/bug-reporting/details

  • Read

    GET /api/cache

    Return cache configuration. Supports pagination via `limit` and `offset` query parameters, and sorting via `sort_column` and `sort_direction`.

    GET /api/cache

  • Read

    GET /api/card

    Get all the Cards. Option filter param `f` can be used to change the set of Cards that are returned; default is `all`, but other options include `mine`, `bookmarked`, `database`, `table`, `using_model`, `using_segment`, and `archived`. See corresponding implementation functions above for the specific behavior of each filter option. :card_index:

    GET /api/card

  • Read

    GET /api/card/embeddable

    Fetch a list of Cards where `enable_embedding` is `true`. The cards can be embedded using the embedding endpoints and a signed JWT.

    GET /api/card/embeddable

  • Read

    GET /api/card/public

    Fetch a list of Cards with public UUIDs. These cards are publicly-accessible *if* public sharing is enabled.

    GET /api/card/public

  • Read

    GET /api/card/{card-id}/params/{param-key}/search/{query}

    Fetch possible values of the parameter whose ID is `:param-key` that contain `:query`. ;; fetch values for Card 1 parameter 'abc' that contain 'Orange'; GET /api/queries/1/params/abc/search/Orange Currently limited to first 1000 results.

    GET /api/card/{card-id}/params/{param-key}/search/{query}

  • Read

    GET /api/card/{card-id}/params/{param-key}/values

    Fetch possible values of the parameter whose ID is `:param-key`. ;; fetch values for Card 1 parameter 'abc' that are possible GET /api/queries/1/params/abc/values

    GET /api/card/{card-id}/params/{param-key}/values

  • Read

    GET /api/card/{id}

    Get `Card` with ID. As of v57, returns the MBQL query (`dataset_query`) as MBQL 5; to return the query as MBQL 4 (aka legacy MBQL) instead, you can specify `?legacy-mbql=true`.

    GET /api/card/{id}

  • Read

    GET /api/card/{id}/dashboards

    Get a list of `{:name ... :id ...}` pairs for all the dashboards this card appears in.

    GET /api/card/{id}/dashboards

  • Read

    GET /api/card/{id}/params/{param-key}/remapping

    Fetch the remapped value for a given value of the parameter with ID `:param-key`. ;; fetch the remapped value for Card 1 parameter 'abc' for value 100 GET /api/queries/1/params/abc/remapping?value=100

    GET /api/card/{id}/params/{param-key}/remapping

  • Read

    GET /api/card/{id}/query_metadata

    Get all of the required query metadata for a card.

    GET /api/card/{id}/query_metadata

  • Read

    GET /api/card/{id}/series

    Fetches a list of compatible series with the card with id `card_id`. - `last_cursor` with value is the id of the last card from the previous page to fetch the next page. - `query` to search card by name. - `exclude_ids` to filter out a list of card ids

    GET /api/card/{id}/series

  • Read

    GET /api/channel

    Get all channels

    GET /api/channel

  • Read

    GET /api/channel/{id}

    Get a channel

    GET /api/channel/{id}

  • Read

    GET /api/cloud-migration

    Get the latest cloud migration, if any.

    GET /api/cloud-migration

  • Read

    GET /api/collection

    Fetch a list of all Collections that the current user has read permissions for (`:can_write` is returned as an additional property of each Collection so you can tell which of these you have write permissions for.) By default, this returns non-archived Collections, but instead you can show archived ones by passing `?archived=true`. By default, admin users will see all collections. To hide other user's collections pass in `?exclude-other-user-collections=true`. If personal-only is `true`, then return only personal collections where `personal_owner_id` is not `nil`.

    GET /api/collection

  • Read

    GET /api/collection/graph

    Fetch a graph of all Collection Permissions.

    GET /api/collection/graph

  • Read

    GET /api/collection/root

    Return the 'Root' Collection object with standard details added

    GET /api/collection/root

  • Read

    GET /api/collection/root/dashboard-question-candidates

    Find cards in the root collection that can be moved into dashboards in the root collection. (Same as the above endpoint, but for the root collection)

    GET /api/collection/root/dashboard-question-candidates

  • Read

    GET /api/collection/root/items

    Fetch objects that the current user should see at their root level. As mentioned elsewhere, the 'Root' Collection doesn't actually exist as a row in the application DB: it's simply a virtual Collection where things with no `collection_id` exist. It does, however, have its own set of Permissions. This endpoint will actually show objects with no `collection_id` for Users that have Root Collection permissions, but for people without Root Collection perms, we'll just show the objects that have an effective location of `/`. This endpoint is intended to power a 'Root Folder View' for the Current User, so regardless you'll see all the top-level objects you're allowed to access. By default, this will show the 'normal' Collections namespace; to view a different Collections namespace, such as `snippets`, you can pass the `?namespace=` parameter. By default, library collections are excluded from the results; to include them, pass `?include-library=true`. Pass `?q=` to filter items by name or last editor. Pass `?include-available-models=true` to include the models that have at least one visible item in the requested scope. Note that this endpoint should return results in a similar shape to `/api/dashboard/:id/items`, so if this is changed, that should too.

    GET /api/collection/root/items

  • Read

    GET /api/collection/root/items/metadata

    Metadata about the Root Collection's items list: the models with at least one visible item plus the item count. Unlike `GET /api/collection/root/items`, the result does not depend on search text; pass that endpoint's other scope params so the metadata describes the list being shown.

    GET /api/collection/root/items/metadata

  • Read

    GET /api/collection/trash

    Fetch the trash collection, as in `/api/collection/:trash-id`

    GET /api/collection/trash

  • Read

    GET /api/collection/tree

    Similar to `GET /`, but returns Collections in a tree structure, e.g. ``` [{:name "A" :below #{:card :dataset} :children [{:name "B"} {:name "C" :here #{:dataset :card} :below #{:dataset :card} :children [{:name "D" :here #{:dataset} :children [{:name "E"}]} {:name "F" :here #{:card} :children [{:name "G"}]}]}]} {:name "H"}] ``` The here and below keys indicate the types of items at this particular level of the tree (here) and in its subtree (below). TODO: for historical reasons this returns Saved Questions AS 'card' AND Models as 'dataset'; we should fix this at some point in the future. By default, looks at the `analytics` (if enabled) and regular (`nil`) namespaces. You can optionally pass a `namespace` argument, or one or many `namespaces`, to specify the particular collection namespaces you wish to look at. For example, `namespaces=analytics&namespaces=` would match the default behavior. When `shallow` is true, takes an optional `collection-id` and returns only the requested collection (or the root, if `collection-id` is `nil`).

    GET /api/collection/tree

  • Read

    GET /api/collection/{id}

    Fetch a specific Collection with standard details added

    GET /api/collection/{id}

  • Read

    GET /api/collection/{id}/dashboard-question-candidates

    Find cards in this collection that can be moved into dashboards in this collection. To be eligible, a card must only appear in one dashboard (which is also in this collection), and must not already be a dashboard question.

    GET /api/collection/{id}/dashboard-question-candidates

  • Read

    GET /api/collection/{id}/items

    Fetch a specific Collection's items with the following options: * `models` - only include objects of a specific set of `models`. If unspecified, returns objects of all models * `archived` - when `true`, return archived objects *instead* of unarchived ones. Defaults to `false`. * `pinned-state` - when `is_pinned`, return pinned objects only. when `is_not_pinned`, return non pinned objects only. when `all`, return everything. By default returns everything. * `q` - filter items by name or last editor. Blank or whitespace-only values are ignored. * `include-available-models` - include the models that have at least one visible item in the requested scope. Note that this endpoint should return results in a similar shape to `/api/dashboard/:id/items`, so if this is changed, that should too.

    GET /api/collection/{id}/items

  • Read

    GET /api/collection/{id}/items/metadata

    Metadata about the collection's items list: the models with at least one visible item plus the item count. Unlike `GET /api/collection/:id/items`, the result does not depend on search text; pass that endpoint's other scope params -- `models`, `archived`, `pinned-state`, `show-dashboard-questions`, `show-exploration-documents` -- so the metadata describes the list being shown.

    GET /api/collection/{id}/items/metadata

  • Read

    GET /api/comment

    Get comments for an entity

    GET /api/comment

  • Read

    GET /api/comment/mentions

    Get a list of entities suitable for mentions. NOTE: only users for now.

    GET /api/comment/mentions

  • Read

    GET /api/dashboard

    This endpoint is currently unused by the Metabase frontend and may be out of date with the rest of the application. It only exists for backwards compatibility and may be removed in the future. Get `Dashboards`. With filter option `f` (default `all`), restrict results as follows: * `all` - Return all Dashboards. * `mine` - Return Dashboards created by the current user. * `archived` - Return Dashboards that have been archived. (By default, these are *excluded*.)

    GET /api/dashboard

  • Read

    GET /api/dashboard/embeddable

    Fetch a list of Dashboards where `enable_embedding` is `true`. The dashboards can be embedded using the embedding endpoints and a signed JWT.

    GET /api/dashboard/embeddable

  • Read

    GET /api/dashboard/params/valid-filter-fields

    Utility endpoint for powering Dashboard UI. Given some set of `filtered` Field IDs (presumably Fields used in parameters) and a set of `filtering` Field IDs that will be used to restrict values of `filtered` Fields, for each `filtered` Field ID return the subset of `filtering` Field IDs that would actually be used in a chain filter query with these Fields. e.g. in a chain filter query like GET /api/dashboard/10/params/PARAM_1/values?PARAM_2=100 Assume `PARAM_1` maps to Field 1 and `PARAM_2` maps to Fields 2 and 3. The underlying MBQL query may or may not filter against Fields 2 and 3, depending on whether an FK relationship that lets us create a join against Field 1 can be found. You can use this endpoint to determine which of those Fields is actually used: GET /api/dashboard/params/valid-filter-fields?filtered=1&filtering=2&filtering=3 ;; -> {1 [2 3]} Results are returned as a map of `filtered` Field ID -> subset of `filtering` Field IDs that would be used in chain filter query

    GET /api/dashboard/params/valid-filter-fields

  • Read

    GET /api/dashboard/public

    Fetch a list of Dashboards with public UUIDs. These dashboards are publicly-accessible *if* public sharing is enabled.

    GET /api/dashboard/public

  • Read

    GET /api/dashboard/{id}

    Get Dashboard with ID.

    GET /api/dashboard/{id}

  • Read

    GET /api/dashboard/{id}/items

    Get Dashboard with ID.

    GET /api/dashboard/{id}/items

  • Read

    GET /api/dashboard/{id}/params/{param-key}/remapping

    Fetch the remapped value for a given value of the parameter with ID `:param-key`. ;; fetch the remapped value for Dashboard 1 parameter 'abc' for value 100 GET /api/dashboard/1/params/abc/remapping?value=100

    GET /api/dashboard/{id}/params/{param-key}/remapping

  • Read

    GET /api/dashboard/{id}/params/{param-key}/search/{query}

    Fetch possible values of the parameter whose ID is `:param-key` that contain `:query`. Optionally restrict these values by passing query parameters like `other-parameter=value` e.g. ;; fetch values for Dashboard 1 parameter 'abc' that contain 'Cam' and are possible when parameter 'def' is set ;; to 100 GET /api/dashboard/1/params/abc/search/Cam?def=100 Currently limited to first 1000 results.

    GET /api/dashboard/{id}/params/{param-key}/search/{query}

  • Read

    GET /api/dashboard/{id}/params/{param-key}/values

    Fetch possible values of the parameter whose ID is `:param-key`. If the values come directly from a query, optionally restrict these values by passing query parameters like `other-parameter=value` e.g. ;; fetch values for Dashboard 1 parameter 'abc' that are possible when parameter 'def' is set to 100 GET /api/dashboard/1/params/abc/values?def=100

    GET /api/dashboard/{id}/params/{param-key}/values

  • Read

    GET /api/dashboard/{id}/query_metadata

    Get all of the required query metadata for the cards on dashboard.

    GET /api/dashboard/{id}/query_metadata

  • Read

    GET /api/database

    Fetch all `Databases`. * `include=tables` means we should hydrate the Tables belonging to each DB. Default: `false`. * `include=schemas` hydrates a `:schemas` list on each DB (only schemas with visible tables the current user can read), matching the per-DB `/api/database/:id/schemas` response. Lets clients avoid N round-trips when populating a database/schema picker. * `saved` means we should include the saved questions virtual database. Default: `false`. * `include_editable_data_model` will only include DBs for which the current user has data model editing permissions. (If `include=tables`, this also applies to the list of tables in each DB). Granting those permissions to non-admins requires Enterprise Edition code and the advanced-permissions feature; without both, this is admin-only. * `exclude_uneditable_details` will only include DBs for which the current user can edit the DB details. Has no effect unless Enterprise Edition code is available and the advanced-permissions feature is enabled. * `include_only_uploadable` will only include DBs into which Metabase can insert new data. * `can-query` will only include DBs for which the current user has query permissions. Default: `false`. Independently of these flags, the implementation of [[metabase.models.interface/to-json]] for `:model/Database` in [[metabase.warehouses.models.database]] uses the implementation of [[metabase.models.interface/can-write?]] for `:model/Database` in [[metabase.warehouses.models.database]] to exclude the `details` field, if the requesting user lacks permission to change the database details.

    GET /api/database

  • Read

    GET /api/database/{id}

    Get a single Database with `id`. Optionally pass `?include=tables` or `?include=tables.fields` to include the Tables belonging to this database, or the Tables and Fields, respectively. If the requestor has write permissions for the DB (i.e. is an admin or has data model permissions), then certain inferred secret values will also be included in the returned details (see [[metabase.secrets.models.secret/expand-db-details-inferred-secret-values]] for full details). Passing include_editable_data_model will only return tables for which the current user has data model editing permissions. Granting data model permissions to non-admins requires Enterprise Edition code and a token with the advanced-permissions feature; without both, this is admin-only. In addition, if the user has no data access for the DB (aka block permissions), it will return only the DB name, ID and tables, with no additional metadata. Independently of these flags, the implementation of [[metabase.models.interface/to-json]] for `:model/Database` in [[metabase.warehouses.models.database]] uses the implementation of [[metabase.models.interface/can-write?]] for `:model/Database` in [[metabase.warehouses.models.database]] to exclude the `details` field, if the requesting user lacks permission to change the database details.

    GET /api/database/{id}

  • Read

    GET /api/database/{id}/autocomplete_suggestions

    Return a list of autocomplete suggestions for a given `prefix`, or `substring`. Should only specify one, but `substring` will have priority if both are present. This is intended for use with the ACE Editor when the User is typing raw SQL. Suggestions include matching `Tables` and `Fields` in this `Database`. Tables are returned in the format `[table_name "Table"]`; When Fields have a semantic_type, they are returned in the format `[field_name "table_name base_type semantic_type"]` When Fields lack a semantic_type, they are returned in the format `[field_name "table_name base_type"]`

    GET /api/database/{id}/autocomplete_suggestions

  • Read

    GET /api/database/{id}/card_autocomplete_suggestions

    Return a list of `Card` autocomplete suggestions for a given `query` in a given `Database`. This is intended for use with the ACE Editor when the User is typing in a template tag for a `Card`, e.g. {{#...}}.

    GET /api/database/{id}/card_autocomplete_suggestions

  • Read

    GET /api/database/{id}/fields

    Get a list of all `Fields` in `Database`.

    GET /api/database/{id}/fields

  • Read

    GET /api/database/{id}/healthcheck

    Reports whether the database can currently connect

    GET /api/database/{id}/healthcheck

  • Read

    GET /api/database/{id}/idfields

    Get a list of all primary key `Fields` for `Database`.

    GET /api/database/{id}/idfields

  • Read

    GET /api/database/{id}/metadata

    Get metadata about a `Database`, including all of its `Tables` and `Fields`. Returns DB, fields, and field values. By default only non-hidden tables are returned. Passing include_hidden=true includes them. Passing include_sensitive_fields=true includes fields with visibility_type sensitive, for tables whose data model the current user can edit. Passing include_editable_data_model will only return tables for which the current user has data model editing permissions. Granting data model permissions to non-admins requires Enterprise Edition code and a token with the advanced-permissions feature; without both, this is admin-only. In addition, if the user has no data access for the DB (aka block permissions), it will return only the DB name, ID and tables, with no additional metadata.

    GET /api/database/{id}/metadata

  • Read

    GET /api/database/{id}/schema

    Return a list of Tables for a Database whose schema is the `schema` query parameter, or `nil` or an empty string when no `schema` is given. Unlike the `/:id/schema/:schema` route, the query parameter supports schema names containing slashes, which are rejected at the HTTP layer when percent-encoded in the URL path (#77353). Optional filters: - `can-query=true` - filter to only tables the user can query

    GET /api/database/{id}/schema

  • Read

    GET /api/database/{id}/schema/{schema}

    Returns a list of Tables for the given Database `id` and `schema`. Schema names containing slashes, backslashes, or percent signs are rejected at the HTTP layer when percent-encoded in the URL path; pass those as the `schema` query parameter of `GET /:id/schema/` instead (#77353). Optional filters: - `can-query=true` - filter to only tables the user can query

    GET /api/database/{id}/schema/{schema}

  • Read

    GET /api/database/{id}/schemas

    Returns a list of all the schemas with tables found for the database `id`. Excludes schemas with no tables. Optional filters: - `can-query=true` - filter to only schemas containing tables the user can query

    GET /api/database/{id}/schemas

  • Read

    GET /api/database/{id}/settings-available

    Get all database-local settings and their availability for the given database.

    GET /api/database/{id}/settings-available

  • Read

    GET /api/database/{id}/syncable_schemas

    Returns a list of all syncable schemas found for the database `id`.

    GET /api/database/{id}/syncable_schemas

  • Read

    GET /api/database/{id}/usage_info

    Get usage info for a database. Returns a map with keys are models and values are the number of entities that use this database.

    GET /api/database/{id}/usage_info

  • Read

    GET /api/database/{virtual-db}/datasets

    Returns a list of all the datasets found for the saved questions virtual database.

    GET /api/database/{virtual-db}/datasets

  • Read

    GET /api/database/{virtual-db}/datasets/{schema}

    Returns a list of Tables for the datasets virtual database.

    GET /api/database/{virtual-db}/datasets/{schema}

  • Read

    GET /api/database/{virtual-db}/metadata

    Endpoint that provides metadata for the Saved Questions 'virtual' database. Used for fooling the frontend and allowing it to treat the Saved Questions virtual DB just like any other database.

    GET /api/database/{virtual-db}/metadata

  • Read

    GET /api/database/{virtual-db}/schema/{schema}

    Returns a list of Tables for the saved questions virtual database.

    GET /api/database/{virtual-db}/schema/{schema}

  • Read

    GET /api/database/{virtual-db}/schemas

    Returns a list of all the schemas found for the saved questions virtual database.

    GET /api/database/{virtual-db}/schemas

  • Read

    GET /api/document

    Gets existing `Documents`.

    GET /api/document

  • Read

    GET /api/document/public

    List all Documents that have public links. Returns a sequence of Documents that have been publicly shared. Each Document includes its `:id`, `:name`, and `:public_uuid`. Documents are only actually accessible via the public endpoint if public sharing is currently enabled. Archived Documents are excluded from the results. This endpoint is used to populate the public links listing in the Admin settings UI. Requires superuser permissions. Public sharing must be enabled via the `enable-public-sharing` setting.

    GET /api/document/public

  • Read

    GET /api/document/{document-id}

    Returns an existing Document by ID.

    GET /api/document/{document-id}

  • Read

    GET /api/ee/advanced-permissions/application/graph

    Fetch a graph of Application Permissions.

    GET /api/ee/advanced-permissions/application/graph

  • Read

    GET /api/ee/advanced-permissions/impersonation

    Fetch a list of all Impersonation policies currently in effect, or a single policy if both `group_id` and `db_id` are provided.

    GET /api/ee/advanced-permissions/impersonation

  • Read

    GET /api/ee/ai-controls/permissions

    List all metabot permissions for all groups, filling in defaults for missing entries.

    GET /api/ee/ai-controls/permissions

  • Read

    GET /api/ee/ai-controls/usage/group

    Get all group-level metabot usage limits.

    GET /api/ee/ai-controls/usage/group

  • Read

    GET /api/ee/ai-controls/usage/group/{group-id}

    Get the metabot usage limit for a specific group. Returns `max_usage: null` if no limit is set.

    GET /api/ee/ai-controls/usage/group/{group-id}

  • Read

    GET /api/ee/ai-controls/usage/instance

    Get the instance-wide metabot usage limit. Returns `max_usage: null` if no limit is set (unlimited).

    GET /api/ee/ai-controls/usage/instance

  • Read

    GET /api/ee/ai-controls/usage/tenant

    Get all tenant-level metabot usage limits.

    GET /api/ee/ai-controls/usage/tenant

  • Read

    GET /api/ee/ai-controls/usage/tenant/{tenant-id}

    Get the metabot usage limit for a specific tenant. Returns `max_usage: null` if no limit is set.

    GET /api/ee/ai-controls/usage/tenant/{tenant-id}

  • Read

    GET /api/ee/audit-app/user/audit-info

    Gets audit info for the current user if he has permissions to access the audit collection. Otherwise return an empty map.

    GET /api/ee/audit-app/user/audit-info

  • Read

    GET /api/ee/billing

    Get billing information. This acts as a proxy between `metabase-billing-info-url` and the client, using the embedding token and signed in user's email to fetch the billing information.

    GET /api/ee/billing

  • Read

    GET /api/ee/cloud-add-ons/addons

    Get addons information from the Metabase Store API.

    GET /api/ee/cloud-add-ons/addons

  • Read

    GET /api/ee/cloud-add-ons/plans

    Get plans information from the Metabase Store API.

    GET /api/ee/cloud-add-ons/plans

  • Read

    GET /api/ee/content-translation/csv

    Provides content translation dictionary in CSV

    GET /api/ee/content-translation/csv

  • Read

    GET /api/ee/content-translation/dictionary

    Fetch the content translation dictionary for authenticated users (auth-based embedding flows).

    GET /api/ee/content-translation/dictionary

  • Read

    GET /api/ee/content-translation/dictionary/{token}

    Fetch the content translation dictionary via a JSON Web Token signed with the `embedding-secret-key`.

    GET /api/ee/content-translation/dictionary/{token}

  • Read

    GET /api/ee/custom-viz-plugin

    List all registered custom visualization plugins.

    GET /api/ee/custom-viz-plugin

  • Read

    GET /api/ee/custom-viz-plugin/list

    List active and enabled custom visualization plugins. Available to any authenticated user. Plugins with version mismatches are included, with soft `warnings` attached. Dev-only plugins are excluded when dev mode is disabled.

    GET /api/ee/custom-viz-plugin/list

  • Read

    GET /api/ee/custom-viz-plugin/sandbox-host

    Serve a minimal HTML document used as the iframe `src` for the near-membrane custom-viz sandbox. The response carries a per-document `Content-Security-Policy` that permits `'unsafe-eval'` only inside this iframe, so the main Metabase document keeps its strict nonce-based CSP. 404s unless the custom-viz kill switch is on and the request passes the fetch-metadata gate (see [[donor-load-allowed?]]).

    GET /api/ee/custom-viz-plugin/sandbox-host

  • Read

    GET /api/ee/custom-viz-plugin/{id}/asset

    Serve the plugin's icon image from its bundle. The asset path is passed as a `path` query parameter (e.g. `?path=icon.svg`) and must match the manifest `icon`. Only the icon is served — plugins do not ship arbitrary assets. In dev mode, proxies from the dev base URL if set.

    GET /api/ee/custom-viz-plugin/{id}/asset

  • Read

    GET /api/ee/custom-viz-plugin/{id}/bundle

    Serve the JS bundle for a plugin from the on-disk cache. Returns application/javascript with ETag and Cache-Control headers. In dev mode, proxies from `dev_bundle_url` if set.

    GET /api/ee/custom-viz-plugin/{id}/bundle

  • Read

    GET /api/ee/custom-viz-plugin/{id}/dev-sse

    Proxy Server-Sent Events from the plugin's dev server. Connects to `{dev_bundle_url}/__sse` and forwards events to the browser. This avoids the need for a CSP exception for the dev server origin. Requires custom viz plugin dev mode to be enabled.

    GET /api/ee/custom-viz-plugin/{id}/dev-sse

  • Read

    GET /api/ee/data-complexity-score/complexity

    Return the most recently stored Data Complexity Score for this instance. Pass `force-recalculation=true` to recompute, persist, and return a fresh score. Superuser-only.

    GET /api/ee/data-complexity-score/complexity

  • Read

    GET /api/ee/data-studio/table/{id}/publishing-info

    Return the latest valid publishing information for a published table, or no content when unavailable.

    GET /api/ee/data-studio/table/{id}/publishing-info

  • Read

    GET /api/ee/dependencies/backfill-status

    Returns whether the dependency backfill has pending work. `complete` is true when there are no stale or outdated entities awaiting processing.

    GET /api/ee/dependencies/backfill-status

  • Read

    GET /api/ee/dependencies/graph

    This endpoint takes an :id and a supported entity :type, and returns a graph of all its upstream dependencies. The graph is represented by a list of :nodes and a list of :edges. Each node has an :id, :type, :data (which depends on the node type), and a map of :dependent_counts per entity type. Each edge is a :model/Dependency.

    GET /api/ee/dependencies/graph

  • Read

    GET /api/ee/dependencies/graph/breaking

    Returns a list of entities that are breaking other entities (sources of errors). These are tables or cards that other entities depend on, where those dependents have validation errors traced back to this source entity. Accepts optional parameters for filtering: - `types`: List of source entity types - only `:card` or `:table` (default: both) - `card-types`: List of card types to include when filtering cards (e.g., `[:question :model :metric]`) - `query`: Search string to filter by name or location - `include-personal-collections`: Controls whether items in personal collections are included (default: false) - `sort-column`: Sort column - `:name`, `:location`, `:dependents-errors`, or `:dependents-with-errors` (default: `:name`) - `sort-direction`: Sort direction - `:asc` or `:desc` (default: `:asc`) - `offset`: Default 0 - `limit`: Default 50 Returns a map with: - `data`: List of breaking source entities - `total`: Total count of matched items - `offset`: Applied offset - `limit`: Applied limit

    GET /api/ee/dependencies/graph/breaking

  • Read

    GET /api/ee/dependencies/graph/broken

    Returns the broken dependents for a specific source entity. These are entities that have validation errors traced back to the specified source. Required parameters: - `id`: The ID of the source entity - `type`: The type of the source entity (card, table) Optional parameters: - `dependent-types`: Dependency types to filter by. Can be single value or array. - `dependent-card-types`: Card types to filter by when dependent-types includes :card. - `include-personal-collections`: Include items in personal collections (default: false) - `sort-column`: Column to sort by - name, location, or view-count (default: name) - `sort-direction`: Sort direction - asc or desc (default: asc)

    GET /api/ee/dependencies/graph/broken

  • Read

    GET /api/ee/dependencies/graph/dependents

    Returns a list of dependents for the specified entity. Required parameters: - `id`: The ID of the entity - `type`: The type of the entity (card, table, dashboard, etc.) Optional parameters: - `dependent-types`: Dependency types to filter by. Can be single value or array. If not provided, returns all types. Example: ?dependent-types=card&dependent-types=dashboard - `dependent-card-types`: Card types to filter by when dependent-types includes :card. Ignored if dependent-types doesn't include :card. Example: ?dependent-card-types=question&dependent-card-types=model - `broken`: Return only broken entities (default: false) - `query`: Search string to filter results by name or location (case-insensitive) - `include-personal-collections`: Include items in personal collections (default: false) - `sort-column`: Column to sort by - name, location, or view-count (default: name) - `sort-direction`: Sort direction - asc or desc (default: asc)

    GET /api/ee/dependencies/graph/dependents

  • Read

    GET /api/ee/dependencies/graph/unreferenced

    Returns a list of all unreferenced items in the instance. An unreferenced item is one that is not a dependency of any other item. Accepts optional parameters for filtering: - `types`: List of entity types to include (e.g., [:card :transform :snippet :dashboard]) - `card-types`: List of card types to include when filtering cards (e.g., [:question :model :metric]) - `query`: Search string to filter by name or location - `include-personal-collections`: Controls whether items in personal collections are included (default: false) - `sort-column`: Sort column - `:name`, `:location`, `:dependents-errors`, or `:dependents-with-errors` (default: `:name`) - `sort-direction`: Sort direction - `:asc` or `:desc` (default: `:asc`) - `offset`: Default 0 - `limit`: Default 50 Returns a map with: - `data`: List of unreferenced items, each with `:id`, `:type`, and `:data` fields - `total`: Total count of matched items - `offset`: Applied offset - `limit`: Applied limit

    GET /api/ee/dependencies/graph/unreferenced

  • Read

    GET /api/ee/embedding-hub/checklist

    Get the setup guide checklist status, indicating which setup steps have been completed.

    GET /api/ee/embedding-hub/checklist

  • Read

    GET /api/ee/erd

    Return an Entity Relationship Diagram (ERD) for tables and their FK relationships. When `table-ids` is provided, those tables are the focal points (and must belong to `database-id`).

    GET /api/ee/erd

  • Read

    GET /api/ee/gsheets/connection

    Check the status of a connection. This endpoint gets polled by FE to determine when to stop showing the setup widget. Returns the gsheets shape, with the attached datawarehouse db id at `:db_id`.

    GET /api/ee/gsheets/connection

  • Read

    GET /api/ee/gsheets/service-account

    Checks to see if service-account is setup or not, delegates to HM only if we haven't set it from a metabase cluster before.

    GET /api/ee/gsheets/service-account

  • Read

    GET /api/ee/library

    Get the Library. If no library exists, it doesn't fail but returns an empty response

    GET /api/ee/library

  • Read

    GET /api/ee/library/tree

    This matches /api/collection/tree but only returns the library collection.

    GET /api/ee/library/tree

  • Read

    GET /api/ee/logs/query_execution/{yyyy-mm}

    Fetch rows for the month specified by `:yyyy-mm` from the query_execution logs table. Must be a superuser.

    GET /api/ee/logs/query_execution/{yyyy-mm}

  • Read

    GET /api/ee/metabot-analytics/conversations

    Return paginated list of AI conversations with summary statistics.

    GET /api/ee/metabot-analytics/conversations

  • Read

    GET /api/ee/metabot-analytics/conversations/{id}

    Return full details for a specific conversation including all messages.

    GET /api/ee/metabot-analytics/conversations/{id}

  • Read

    GET /api/ee/metabot/usage

    Fetch current Metabot token usage for the current billing period.

    GET /api/ee/metabot/usage

  • Read

    GET /api/ee/mfa/admin/enrolled-users

    Admin: users who have a confirmed second factor. Never feature-gated, for the same reason `/admin/remove` isn't — after a licence lapse an admin must still be able to find and unlock a locked-out user. Takes `limit`/`offset` for pagination, and `query` to search on first name, last name, and email.

    GET /api/ee/mfa/admin/enrolled-users

  • Read

    GET /api/ee/mfa/admin/overview

    Admin: enrollment overview — how many users have (and haven't) set up a second factor, and whether the instance encrypts secrets at rest.

    GET /api/ee/mfa/admin/overview

  • Read

    GET /api/ee/mfa/admin/unenrolled-users

    Admin: active users who have not set up a second factor. Matches `unenrolled_count` from `/admin/overview`, so people who sign in through SSO are included even though the login gate never challenges them — their identity provider owns MFA. Never feature-gated, as above. Takes `limit`/`offset` for pagination, and `query` to search on first name, last name, and email.

    GET /api/ee/mfa/admin/unenrolled-users

  • Read

    GET /api/ee/mfa/status

    The current user's MFA status, for the account-settings UI.

    GET /api/ee/mfa/status

  • Read

    GET /api/ee/permission_debug

    This endpoint expects a `user_id`, a `model_id` to debug permissions against, and `action_type`. The type of model we are debugging against is inferred by the `action_type`. It will return: - `decision`: The overall permission decision ("allow", "denied", or "limited") - `model-type`: The type of model being checked (e.g., "question") - `model-id`: The ID of the model being checked - `segment`: A set of segmentation types applied (e.g., "sandboxed", "impersonated", "routed") - `message`: A sequence of strings explaining the decision - `data`: A map containing details about permissions (table or collection names to group names) - `suggestions`: A map of group IDs to group names that could provide access Example requests: - Check if user can read a card: `GET /api/ee/permission_debug?user_id=123&model_id=456&action_type=card/read` - Check if user can query a card: `GET /api/ee/permission_debug?user_id=123&model_id=456&action_type=card/query` - Check if user can download data: `GET /api/ee/permission_debug?user_id=123&model_id=456&action_type=card/download-data` Example responses: - Allowed access: ```json { "decision": "allow", "model-type": "question", "model-id": "456", "segment": [], "message": ["User has permission to read this card"], "data": {}, "suggestions": {} } ``` - Denied access with blocked table: ```json { "decision": "denied", "model-type": "question", "model-id": "456", "segment": [], "message": ["User does not have permission to query this card"], "data": {"sample-db.PUBLIC.ORDERS": ["All Users"]}, "suggestions": {} } ``` - Limited access: ```json { "decision": "limited", "model-type": "question", "model-id": "456", "segment": [], "message": ["User has permission to download some data from this card"], "data": {}, "suggestions": {} } ```

    GET /api/ee/permission_debug

  • Read

    GET /api/ee/remote-sync/branches

    Get list of branches from the configured source. Returns a JSON object with branch names under the :items key. Requires superuser permissions.

    GET /api/ee/remote-sync/branches

  • Read

    GET /api/ee/remote-sync/current-task

    Get the current sync task

    GET /api/ee/remote-sync/current-task

  • Read

    GET /api/ee/remote-sync/dirty

    Return all models with changes that have not been pushed to the remote sync source in any remote-synced collection.

    GET /api/ee/remote-sync/dirty

  • Read

    GET /api/ee/remote-sync/export-preflight

    Dry-run preview of what pushing the current state would do given the live remote branch, without writing anything. Drives the UI's push decision (force / new branch / merge). Returns: - has_changes: whether the remote branch has advanced beyond the last synced version - clean: whether a 3-way merge would apply with no conflicts - conflicts: human-readable labels of the entities that conflict (empty when clean) - summary: counts of remote changes a merge would fold in - force_push_casualties: remote content a force push (rather than a merge) would discard, as `{:deleted [labels] :overwritten [labels]}` - reason: "history-rewritten" when the remote was force-pushed/rebased so no merge base exists Requires superuser permissions.

    GET /api/ee/remote-sync/export-preflight

  • Read

    GET /api/ee/remote-sync/has-remote-changes

    Check if there are new changes on the remote branch that can be pulled. Uses in-memory caching (configurable TTL via remote-sync-check-changes-cache-ttl-seconds setting). Returns: - has_changes: true if remote version differs from last imported version, or if never imported - remote_version: current Git SHA on remote branch - local_version: Git SHA of last successful import (nil if never imported) - cached: true if result was served from cache

    GET /api/ee/remote-sync/has-remote-changes

  • Read

    GET /api/ee/remote-sync/is-dirty

    Check if any remote-synced collection or collection item has local changes that have not been pushed to the remote sync source.

    GET /api/ee/remote-sync/is-dirty

  • Read

    GET /api/ee/replacement/runs

    List replacement runs, optionally filtered by is-active.

    GET /api/ee/replacement/runs

  • Read

    GET /api/ee/replacement/runs/{id}

    Get the status of a source replacement run.

    GET /api/ee/replacement/runs/{id}

  • Read

    GET /api/ee/scim/api_key

    Fetch the SCIM API key if one exists. Does *not* return an unmasked key, since we don't have access to that after it is created.

    GET /api/ee/scim/api_key

  • Read

    GET /api/ee/scim/v2/Groups

    Fetch a list of groups.

    GET /api/ee/scim/v2/Groups

  • Read

    GET /api/ee/scim/v2/Groups/{id}

    Fetch a single group.

    GET /api/ee/scim/v2/Groups/{id}

  • Read

    GET /api/ee/scim/v2/Users

    Fetch a list of users.

    GET /api/ee/scim/v2/Users

  • Read

    GET /api/ee/scim/v2/Users/{id}

    Fetch a single user.

    GET /api/ee/scim/v2/Users/{id}

  • Read

    GET /api/ee/security-center

    List all security advisories with match status.

    GET /api/ee/security-center

  • Read

    GET /api/ee/semantic-search/status

    Fetch the indexing status of the currently active semantic search index table. Returns a map with keys: :indexed_count <number of indexed items> :total_est <estimated total number of items to index> If no index is active, or no pgvector database is configured, returns an empty map.

    GET /api/ee/semantic-search/status

  • Read

    GET /api/ee/stale/{id}

    A flexible endpoint that returns stale entities, in the same shape as collections/items, with the following options: - `before_date` - only return entities that were last edited before this date (default: 6 months ago) - `is_recursive` - if true, return entities from all children of the collection, not just the direct children (default: false) - `sort_column` - the column to sort by (default: name) - `sort_direction` - the direction to sort by (default: asc)

    GET /api/ee/stale/{id}

  • Read

    GET /api/ee/support-access-grant

    List support access grants with optional filtering and pagination. Requires superuser permissions. Query parameters: - ticket-number: Filter by ticket number - user-id: Filter by user who created the grant - include-revoked: Include revoked grants (default false)

    GET /api/ee/support-access-grant

  • Read

    GET /api/ee/support-access-grant/current

    Get the currently active support access grant, if one exists. Requires superuser permissions.

    GET /api/ee/support-access-grant/current

  • Read

    GET /api/ee/tenant

    Get all tenants

    GET /api/ee/tenant

  • Read

    GET /api/ee/tenant/{id}

    Get info about a tenant

    GET /api/ee/tenant/{id}

  • Read

    GET /api/ee/transforms-python/library/{path}

    Get the Python library for user modules.

    GET /api/ee/transforms-python/library/{path}

  • Read

    GET /api/ee/transforms/{id}/inspect

    Phase 1: Discover available lenses for a transform. Returns structural metadata and available lens types.

    GET /api/ee/transforms/{id}/inspect

  • Read

    GET /api/ee/transforms/{id}/inspect/{lens-id}

    Phase 2: Get full lens contents for a transform. Returns sections, cards with dataset_query, and trigger definitions. Accepts optional params for drill lenses as query params.

    GET /api/ee/transforms/{id}/inspect/{lens-id}

  • Read

    GET /api/ee/upload-management/tables

    Get all `Tables` visible to the current user which were created by uploading a file.

    GET /api/ee/upload-management/tables

  • Read

    GET /api/embed-mcp/bootstrap

    Everything the MCP Apps iframe needs to mount, so its UI credential never authenticates a general REST endpoint. Replaces the iframe's calls to `GET /api/user/current` and `GET /api/session/properties`.

    GET /api/embed-mcp/bootstrap

  • Read

    GET /api/embed-mcp/queries/{handle}

    Resolve a query handle to the base64-encoded MBQL the iframe should render. This is how the v2 MCP Apps tools keep result data out of the model context: `visualize_query` and `render_drill_through` return only a handle, and the iframe exchanges it here using the scoped UI credential it was rendered with. The lookup is user-scoped (see [[metabase.mcp.session/find-handle-row]]) and the credential is accepted only on the MCP UI request surface, so a handle on its own is not a bearer credential.

    GET /api/embed-mcp/queries/{handle}

  • Read

    GET /api/embed-theme

    Fetch a list of all embedding themes.

    GET /api/embed-theme

  • Read

    GET /api/embed-theme/{id}

    Fetch a single embedding theme by ID.

    GET /api/embed-theme/{id}

  • Read

    GET /api/embed/card/{token}

    Fetch a Card via a JSON Web Token signed with the `embedding-secret-key`. Token should have the following format: {:resource {:question <card-id>}}

    GET /api/embed/card/{token}

  • Read

    GET /api/embed/card/{token}/params/{param-key}/remapping

    Embedded version of api.card filter values endpoint.

    GET /api/embed/card/{token}/params/{param-key}/remapping

  • Read

    GET /api/embed/card/{token}/params/{param-key}/search/{prefix}

    Embedded version of chain filter search endpoint.

    GET /api/embed/card/{token}/params/{param-key}/search/{prefix}

  • Read

    GET /api/embed/card/{token}/params/{param-key}/values

    Embedded version of api.card filter values endpoint.

    GET /api/embed/card/{token}/params/{param-key}/values

  • Read

    GET /api/embed/card/{token}/query

    Fetch the results of running a Card using a JSON Web Token signed with the `embedding-secret-key`. Token should have the following format: {:resource {:question <card-id>} :params <parameters>}

    GET /api/embed/card/{token}/query

  • Read

    GET /api/embed/card/{token}/query/{export-format}

    Like `GET /api/embed/card/query`, but returns the results as a file in the specified format.

    GET /api/embed/card/{token}/query/{export-format}

  • Read

    GET /api/embed/dashboard/{token}

    Fetch a Dashboard via a JSON Web Token signed with the `embedding-secret-key`. Token should have the following format: {:resource {:dashboard <dashboard-id>}}

    GET /api/embed/dashboard/{token}

  • Read

    GET /api/embed/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}

    Fetch the results of running a Card belonging to a Dashboard using a JSON Web Token signed with the `embedding-secret-key`

    GET /api/embed/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}

  • Read

    GET /api/embed/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}/{export-format}

    Fetch the results of running a Card belonging to a Dashboard using a JSON Web Token signed with the `embedding-secret-key` return the data in one of the export formats

    GET /api/embed/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}/{export-format}

  • Read

    GET /api/embed/dashboard/{token}/params/{param-key}/remapping

    Embedded version of the remapped dashboard param value endpoint.

    GET /api/embed/dashboard/{token}/params/{param-key}/remapping

  • Read

    GET /api/embed/dashboard/{token}/params/{param-key}/search/{prefix}

    Embedded version of chain filter search endpoint.

    GET /api/embed/dashboard/{token}/params/{param-key}/search/{prefix}

  • Read

    GET /api/embed/dashboard/{token}/params/{param-key}/values

    Embedded version of chain filter values endpoint.

    GET /api/embed/dashboard/{token}/params/{param-key}/values

  • Read

    GET /api/embed/pivot/card/{token}/query

    Fetch the results of running a Card using a JSON Web Token signed with the `embedding-secret-key`. Token should have the following format: {:resource {:question <card-id>} :params <parameters>}

    GET /api/embed/pivot/card/{token}/query

  • Read

    GET /api/embed/pivot/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}

    Fetch the results of running a Card belonging to a Dashboard using a JSON Web Token signed with the `embedding-secret-key`

    GET /api/embed/pivot/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}

  • Read

    GET /api/embed/tiles/card/{token}/{zoom}/{x}/{y}

    Generates a single tile image for an embedded Card using the map visualization.

    GET /api/embed/tiles/card/{token}/{zoom}/{x}/{y}

  • Read

    GET /api/embed/tiles/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}/{zoom}/{x}/{y}

    Generates a single tile image for a Card on an embedded Dashboard using the map visualization.

    GET /api/embed/tiles/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}/{zoom}/{x}/{y}

  • Read

    GET /api/embedding-sdk/react-18-hosts

    The host apps that used the React embedding SDK with React 18 in the last 4 days, newest first, at most 50. `host` is null when the host app page URL has no hostname.

    GET /api/embedding-sdk/react-18-hosts

  • Read

    GET /api/eval-trace/{session-id}

    Return the captured eval trace as JSONL (one span per line) for `session-id`. Superuser-only; served only when `MB_AI_EVAL_CAPTURE` is enabled.

    GET /api/eval-trace/{session-id}

  • Read

    GET /api/field/{id}

    Get `Field` with ID.

    GET /api/field/{id}

  • Read

    GET /api/field/{id}/remapping/{remapped-id}

    Fetch remapped Field values.

    GET /api/field/{id}/remapping/{remapped-id}

  • Read

    GET /api/field/{id}/search/{search-id}

    Search for values of a Field with `search-id` that start with `value`. See docstring for [[metabase.parameters.field/search-values]] for a more detailed explanation.

    GET /api/field/{id}/search/{search-id}

  • Read

    GET /api/field/{id}/summary

    Get the count and distinct count of `Field` with ID.

    GET /api/field/{id}/summary

  • Read

    GET /api/field/{id}/values

    If a Field's value of `has_field_values` is `:list`, return a list of all the distinct values of the Field (or remapped Field), and (if defined by a User) a map of human-readable remapped values. If `has_field_values` is not `:list`, checks whether we should create FieldValues for this Field; if so, creates and returns them.

    GET /api/field/{id}/values

  • Read

    GET /api/geojson

    Load a custom GeoJSON file based on a URL or file path provided as a query parameter. This behaves similarly to /api/geojson/:key but doesn't require the custom map to be saved to the DB first.

    GET /api/geojson

  • Read

    GET /api/geojson/{key}

    Fetch a custom GeoJSON file as defined in the [[metabase.geojson.settings/custom-geojson]] setting. (This just acts as a simple proxy for the file specified for `key`).

    GET /api/geojson/{key}

  • Read

    GET /api/glossary

    Fetch all glossary entries, optionally filtered by search term.

    GET /api/glossary

  • Read

    GET /api/health-inspector

    Get a list of recent health check runs.

    GET /api/health-inspector

  • Read

    GET /api/index

    A transform's indexes: those physically in the warehouse, merged with its managed requests. Each entry is flagged `:metabase_managed`; managed ones also carry `:request` (status + definition). If the warehouse can't be read, the managed requests are still listed (as absent) and `:warehouse_error` carries the reason.

    GET /api/index

  • Read

    GET /api/index/request/{id}

    Fetch a single index request (e.g. to poll its status).

    GET /api/index/request/{id}

  • Read

    GET /api/llm/models

    List the models available from every configured connection, grouped by connection. A connection whose credentials are rejected comes back with an empty model list and an `error`, so one bad connection does not blank out the others.

    GET /api/llm/models

  • Read

    GET /api/llm/provider-types

    List the provider types a connection can be created for, with the credential fields each one needs.

    GET /api/llm/provider-types

  • Read

    GET /api/llm/providers

    List the configured provider connections, with their secrets masked.

    GET /api/llm/providers

  • Read

    GET /api/logger/logs

    Logs.

    GET /api/logger/logs

  • Read

    GET /api/logger/presets

    Get all known presets.

    GET /api/logger/presets

  • Read

    GET /api/login-history/current

    Fetch recent logins for the current user.

    GET /api/login-history/current

  • Read

    GET /api/measure

    Fetch *all* `Measures`.

    GET /api/measure

  • Read

    GET /api/measure/{id}

    Fetch `Measure` with ID.

    GET /api/measure/{id}

  • Read

    GET /api/measure/{id}/dimension/{dimension-key}/remapping

    Fetch remapped value for a specific dimension value. Returns a pair [value, display-name] if remapping exists, or [value] otherwise.

    GET /api/measure/{id}/dimension/{dimension-key}/remapping

  • Read

    GET /api/measure/{id}/dimension/{dimension-key}/values

    Fetch values for a dimension of a measure. Returns field values in the same format as the field values API: - values: list of [value] or [value, display-name] tuples - field_id: the underlying field ID - has_more_values: boolean indicating if there are more values

    GET /api/measure/{id}/dimension/{dimension-key}/values

  • Read

    GET /api/metabot/conversations

    List conversations visible in the current user's history, most-recent first. New conversations are participation-based (the user authored at least one message); legacy conversations created before message authors were stamped fall back to the conversation originator. Optionally filter by the last live message's `profile_id`.

    GET /api/metabot/conversations

  • Read

    GET /api/metabot/conversations/{id}

    Return a single conversation with its flattened chat messages. Accessible to any participant in the conversation or to any superuser.

    GET /api/metabot/conversations/{id}

  • Read

    GET /api/metabot/conversations/{id}/title

    Return the title generation status for a conversation. Accessible to any participant in the conversation or to any superuser.

    GET /api/metabot/conversations/{id}/title

  • Read

    GET /api/metabot/metabot

    List configured metabot instances

    GET /api/metabot/metabot

  • Read

    GET /api/metabot/metabot/{id}

    Retrieve one metabot instance

    GET /api/metabot/metabot/{id}

  • Read

    GET /api/metabot/metabot/{id}/prompt-suggestions

    Return the prompt suggestions for the metabot instance with `id`.

    GET /api/metabot/metabot/{id}/prompt-suggestions

  • Read

    GET /api/metabot/permissions/user-permissions

    Return the current user's resolved metabot permissions, taking the most permissive value across all their groups.

    GET /api/metabot/permissions/user-permissions

  • Read

    GET /api/metric

    Get a list of metrics. Returns metrics (Cards with type='metric') that the current user has read access to, filtered by collection visibility permissions.

    GET /api/metric

  • Read

    GET /api/metric/{id}

    Fetch a `Metric` with ID. Returns the metric with hydrated dimensions and dimension mappings; the `::MetricWithDimensions` schema encodes them to the snake_case wire shape at the `defendpoint` edge.

    GET /api/metric/{id}

  • Read

    GET /api/metric/{id}/dimension

    List a metric's curated dimensions, and (when `with-addable=true`) the columns still available to add, grouped by source table. Both can be filtered with a `query` name substring.

    GET /api/metric/{id}/dimension

  • Read

    GET /api/metric/{id}/dimension/{dimension-key}/remapping

    Fetch remapped value for a specific dimension value. Returns a pair [value, display-name] if remapping exists, or [value] otherwise.

    GET /api/metric/{id}/dimension/{dimension-key}/remapping

  • Read

    GET /api/metric/{id}/dimension/{dimension-key}/values

    Fetch values for a dimension of a metric. Returns field values in the same format as the field values API: - values: list of [value] or [value, display-name] tuples - field_id: the underlying field ID - has_more_values: boolean indicating if there are more values

    GET /api/metric/{id}/dimension/{dimension-key}/values

  • Read

    GET /api/model-index

    Retrieve list of ModelIndex.

    GET /api/model-index

  • Read

    GET /api/model-index/{id}

    Retrieve ModelIndex.

    GET /api/model-index/{id}

  • Read

    GET /api/mt/gtap

    Fetch a list of all GTAPs currently in use, or a single GTAP if both `group_id` and `table_id` are provided.

    GET /api/mt/gtap

  • Read

    GET /api/mt/gtap/{id}

    Fetch GTAP by `id`

    GET /api/mt/gtap/{id}

  • Read

    GET /api/mt/user/attributes

    Fetch a list of possible keys for User `login_attributes`. This includes keys from tenant model attributes and keys that have already been set for existing Users.

    GET /api/mt/user/attributes

  • Read

    GET /api/native-query-snippet

    Fetch all snippets

    GET /api/native-query-snippet

  • Read

    GET /api/native-query-snippet/{id}

    Fetch native query snippet with ID.

    GET /api/native-query-snippet/{id}

  • Read

    GET /api/notification

    List notifications. - `creator_id`: if provided returns only notifications created by this user - `recipient_id`: if provided returns only notification that has recipient_id as a recipient - `creator_or_recipient_id`: if provided returns only notification that has user_id as creator or recipient - `card_id`: if provided returns only notification that has card_id as payload

    GET /api/notification

  • Read

    GET /api/notification/admin

    List card-type notifications (alerts) for admin management. Supports pagination (`limit` + `offset` query params — handled by the offset-paging middleware), filtering, and sorting. `last_send_status` filter operates on the latest channel-send task_history row for the notification (`successful` = latest channel-send succeeded; `failing` = latest channel-send failed). `last_check_status` filter operates on the latest terminal TaskRun for the notification — the whole-run rollup (`successful` = the run succeeded; `failing` = it failed or was abandoned). This is a superset of `last_send_status`: it also catches query failures and heartbeat-abandoned runs that never reached the send step. The Failing tab uses `?last_check_status=failing`. `creatorless=true` selects notifications with no creator or a deactivated creator (the Ownerless tab). `creatorless=false` selects the inverse. `channel` accepts a single string or a repeated query param for multi-select (OR logic).

    GET /api/notification/admin

  • Read

    GET /api/notification/admin/{id}

    Get a single card-type notification with last_check, last_send, check_history (up to 10 most-recent terminal alert-type TaskRuns) and send_history (up to 10 most-recent channel-send delivery attempts). 404 if the notification doesn't exist or isn't a card-type notification.

    GET /api/notification/admin/{id}

  • Read

    GET /api/notification/{id}

    Get a notification by id.

    GET /api/notification/{id}

  • Read

    GET /api/oauth/authorizations

    List OAuth dynamic client registration events (registered, approved, or denied), newest first. Joins client info and, for decision events, the deciding user. Superuser only.

    GET /api/oauth/authorizations

  • Read

    GET /api/osi/ai-context

    Get all ai_context entries, paginated.

    GET /api/osi/ai-context

  • Read

    GET /api/osi/ai-context/{entity-type}/{entity-local-id}

    Get the ai_context entry for an entity by its logical `(entity_type, entity_local_id)` key.

    GET /api/osi/ai-context/{entity-type}/{entity-local-id}

  • Read

    GET /api/permissions/graph

    Fetch a graph of all Permissions.

    GET /api/permissions/graph

  • Read

    GET /api/permissions/graph/db/{db-id}

    Fetch a graph of all Permissions for db-id `db-id`.

    GET /api/permissions/graph/db/{db-id}

  • Read

    GET /api/permissions/graph/group/{group-id}

    Fetch a graph of all Permissions for group-id `group-id`.

    GET /api/permissions/graph/group/{group-id}

  • Read

    GET /api/permissions/group

    Fetch all `PermissionsGroups`, including a count of the number of `:members` in that group. This API requires superuser or group manager of more than one group. Group manager is only available if `advanced-permissions` is enabled and returns only groups that user is manager of. Optional query parameter `tenancy` can be used to filter groups: - `tenancy=external`: Returns only tenant groups (where `is_tenant_group = true`) - `tenancy=internal`: Returns only non-tenant groups (where `is_tenant_group = false`) - No `tenancy` parameter: Returns all groups (default behavior)

    GET /api/permissions/group

  • Read

    GET /api/permissions/group/{id}

    Fetch the details for a certain permissions group.

    GET /api/permissions/group/{id}

  • Read

    GET /api/permissions/invite-group-ids

    IDs of the permission groups holding a stored read (or read-write) grant on the collection of a shareable item (a `dashboard` or a `question`). The "invite someone to view" group picker lists all groups and uses these ids to mark the ones whose members can already see the item; the Administrators group has implicit access to everything and is never included. The ids are otherwise unfiltered; system-managed groups like Data Analysts appear when they hold a grant, and clients intersect the ids with the groups they display. Superuser-only, like the invite action itself.

    GET /api/permissions/invite-group-ids

  • Read

    GET /api/permissions/membership

    Fetch a map describing the group memberships of various users. This map's format is: {<user-id> [{:membership_id <id> :group_id <id> :is_group_manager boolean}]}

    GET /api/permissions/membership

  • Read

    GET /api/persist

    List the entries of [[PersistedInfo]] in order to show a status page.

    GET /api/persist

  • Read

    GET /api/persist/card/{card-id}

    Fetch a particular [[PersistedInfo]] by card-id.

    GET /api/persist/card/{card-id}

  • Read

    GET /api/persist/{persisted-info-id}

    Fetch a particular [[PersistedInfo]] by id.

    GET /api/persist/{persisted-info-id}

  • Read

    GET /api/premium-features/token/status

    Fetch info about the current Premium-Features premium features token including whether it is `valid`, a `trial` token, its `features`, when it is `valid-thru`, and the `status` of the account.

    GET /api/premium-features/token/status

  • Read

    GET /api/preview_embed/card/{token}

    Fetch a Card you're considering embedding by passing a JWT `token`.

    GET /api/preview_embed/card/{token}

  • Read

    GET /api/preview_embed/card/{token}/params/{param-key}/remapping

    Embedded version of api.card filter values endpoint.

    GET /api/preview_embed/card/{token}/params/{param-key}/remapping

  • Read

    GET /api/preview_embed/card/{token}/params/{param-key}/values

    Embedded version of api.card filter values endpoint.

    GET /api/preview_embed/card/{token}/params/{param-key}/values

  • Read

    GET /api/preview_embed/card/{token}/query

    Fetch the query results for a Card you're considering embedding by passing a JWT `token`.

    GET /api/preview_embed/card/{token}/query

  • Read

    GET /api/preview_embed/dashboard/{token}

    Fetch a Dashboard you're considering embedding by passing a JWT `token`.

    GET /api/preview_embed/dashboard/{token}

  • Read

    GET /api/preview_embed/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}

    Fetch the results of running a Card belonging to a Dashboard you're considering embedding with JWT `token`.

    GET /api/preview_embed/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}

  • Read

    GET /api/preview_embed/dashboard/{token}/params/{param-key}/remapping

    Embedded version of the remapped dashboard param value endpoint.

    GET /api/preview_embed/dashboard/{token}/params/{param-key}/remapping

  • Read

    GET /api/preview_embed/dashboard/{token}/params/{param-key}/search/{prefix}

    Embedded version of chain filter search endpoint.

    GET /api/preview_embed/dashboard/{token}/params/{param-key}/search/{prefix}

  • Read

    GET /api/preview_embed/dashboard/{token}/params/{param-key}/values

    Embedded version of chain filter values endpoint.

    GET /api/preview_embed/dashboard/{token}/params/{param-key}/values

  • Read

    GET /api/preview_embed/pivot/card/{token}/query

    Fetch the query results for a Card you're considering embedding by passing a JWT `token`.

    GET /api/preview_embed/pivot/card/{token}/query

  • Read

    GET /api/preview_embed/pivot/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}

    Fetch the results of running a Card belonging to a Dashboard you're considering embedding with JWT `token`.

    GET /api/preview_embed/pivot/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}

  • Read

    GET /api/preview_embed/tiles/card/{token}/{zoom}/{x}/{y}

    Generates a single tile image for an embedded Card using the map visualization.

    GET /api/preview_embed/tiles/card/{token}/{zoom}/{x}/{y}

  • Read

    GET /api/preview_embed/tiles/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}/{zoom}/{x}/{y}

    Generates a single tile image for a Card on an embedded Dashboard using the map visualization.

    GET /api/preview_embed/tiles/dashboard/{token}/dashcard/{dashcard-id}/card/{card-id}/{zoom}/{x}/{y}

  • Read

    GET /api/public/action/{uuid}

    Fetch a publicly-accessible Action. Does not require auth credentials. Public sharing must be enabled.

    GET /api/public/action/{uuid}

  • Read

    GET /api/public/card/{uuid}

    Fetch a publicly-accessible Card an return query results as well as `:card` information. Does not require auth credentials. Public sharing must be enabled.

    GET /api/public/card/{uuid}

  • Read

    GET /api/public/card/{uuid}/params/{param-key}/remapping

    Fetch the remapped value for the given `value` of parameter with ID `:param-key` of card with UUID `uuid`.

    GET /api/public/card/{uuid}/params/{param-key}/remapping

  • Read

    GET /api/public/card/{uuid}/params/{param-key}/search/{query}

    Fetch values for a parameter on a public card containing `query`.

    GET /api/public/card/{uuid}/params/{param-key}/search/{query}

  • Read

    GET /api/public/card/{uuid}/params/{param-key}/values

    Fetch values for a parameter on a public card.

    GET /api/public/card/{uuid}/params/{param-key}/values

  • Read

    GET /api/public/card/{uuid}/query

    Fetch a publicly-accessible Card an return query results as well as `:card` information. Does not require auth credentials. Public sharing must be enabled.

    GET /api/public/card/{uuid}/query

  • Read

    GET /api/public/card/{uuid}/query/{export-format}

    Fetch a publicly-accessible Card and return query results in the specified format. Does not require auth credentials. Public sharing must be enabled.

    GET /api/public/card/{uuid}/query/{export-format}

  • Read

    GET /api/public/dashboard/{uuid}

    Fetch a publicly-accessible Dashboard. Does not require auth credentials. Public sharing must be enabled.

    GET /api/public/dashboard/{uuid}

  • Read

    GET /api/public/dashboard/{uuid}/dashcard/{dashcard-id}/card/{card-id}

    Fetch the results for a Card in a publicly-accessible Dashboard. Does not require auth credentials. Public sharing must be enabled.

    GET /api/public/dashboard/{uuid}/dashcard/{dashcard-id}/card/{card-id}

  • Read

    GET /api/public/dashboard/{uuid}/params/{param-key}/remapping

    Fetch the remapped value for the given `value` of parameter with ID `:param-key` of dashboard with UUID `uuid`.

    GET /api/public/dashboard/{uuid}/params/{param-key}/remapping

  • Read

    GET /api/public/dashboard/{uuid}/params/{param-key}/search/{query}

    Fetch filter values for dashboard parameter `param-key`, containing specified `query`.

    GET /api/public/dashboard/{uuid}/params/{param-key}/search/{query}

  • Read

    GET /api/public/dashboard/{uuid}/params/{param-key}/values

    Fetch filter values for dashboard parameter `param-key`.

    GET /api/public/dashboard/{uuid}/params/{param-key}/values

  • Read

    GET /api/public/document/{uuid}

    Fetch a publicly-accessible Document. Does not require auth credentials. Public sharing must be enabled. Returns a Document with sensitive fields removed (excludes collection_id, permissions, creator details, etc.). Includes all embedded Cards with their metadata hydrated so the frontend doesn't need to make separate requests for each card — just like public Dashboards do.

    GET /api/public/document/{uuid}

  • Read

    GET /api/public/document/{uuid}/card/{card-id}

    Run a query for a Card that's embedded in a public Document. Doesn't require auth credentials. Public sharing must be enabled.

    GET /api/public/document/{uuid}/card/{card-id}

  • Read

    GET /api/public/oembed

    oEmbed endpoint used to retrieve embed code and metadata for a (public) Metabase URL.

    GET /api/public/oembed

  • Read

    GET /api/public/pivot/card/{uuid}/query

    Fetch a publicly-accessible Card an return query results as well as `:card` information. Does not require auth credentials. Public sharing must be enabled.

    GET /api/public/pivot/card/{uuid}/query

  • Read

    GET /api/public/pivot/dashboard/{uuid}/dashcard/{dashcard-id}/card/{card-id}

    Fetch the results for a Card in a publicly-accessible Dashboard. Does not require auth credentials. Public sharing must be enabled.

    GET /api/public/pivot/dashboard/{uuid}/dashcard/{dashcard-id}/card/{card-id}

  • Read

    GET /api/public/tiles/card/{uuid}/{zoom}/{x}/{y}

    Generates a single tile image for a publicly-accessible Card using the map visualization. Does not require auth credentials. Public sharing must be enabled.

    GET /api/public/tiles/card/{uuid}/{zoom}/{x}/{y}

  • Read

    GET /api/public/tiles/dashboard/{uuid}/dashcard/{dashcard-id}/card/{card-id}/{zoom}/{x}/{y}

    Generates a single tile image for a Card using the map visualization in a publicly-accessible Dashboard. Does not require auth credentials. Public sharing must be enabled.

    GET /api/public/tiles/dashboard/{uuid}/dashcard/{dashcard-id}/card/{card-id}/{zoom}/{x}/{y}

  • Read

    GET /api/pulse

    Fetch all dashboard subscriptions. By default, returns only subscriptions for which the current user has write permissions. For admins, this is all subscriptions; for non-admins, it is only subscriptions that they created. If `dashboard_id` is specified, restricts results to subscriptions for that dashboard. If `created_or_receive` is `true`, it specifically returns all subscriptions for which the current user created *or* is a known recipient of. Note that this is a superset of the default items returned for non-admins, and a subset of the default items returned for admins. This is used to power the /account/notifications page. This may include subscriptions which the current user does not have collection permissions for, in which case some sensitive metadata (the list of cards and recipients) is stripped out.

    GET /api/pulse

  • Read

    GET /api/pulse/form_input

    Provides relevant configuration information and user choices for creating/updating Pulses.

    GET /api/pulse/form_input

  • Read

    GET /api/pulse/{id}

    Fetch `Pulse` with ID. If the user is a recipient of the Pulse but does not have read permissions for its collection, we still return it but with some sensitive metadata removed.

    GET /api/pulse/{id}

  • Read

    GET /api/revision

    Get revisions of an object.

    GET /api/revision

  • Read

    GET /api/revision/{entity}/{id}

    Fetch `Revisions` for an object with ID.

    GET /api/revision/{entity}/{id}

  • Read

    GET /api/search/debug

    Superuser-only. Explain why `expected_result_id` / `expected_result_type` does not appear in the results of the given search query. Accepts every `GET /api/search` parameter plus the two expected-result parameters, and returns `{:type ..., :details ...}` for the first stage that drops the item: `not-searchable`, `missing-from-index`, `not-permitted`, `filtered`, `not-matching`, or the terminal `matched` / `ranked-out`. - `for_user_id`: diagnose from another user's perspective (defaults to you). Permission and visibility checks run as that user, so a `not-permitted` result means *they* can't see the item. Not supported for `indexed-entity` (its id is compound).

    GET /api/search/debug

  • Read

    GET /api/search/weights

    Return the current weights being used to rank the search results

    GET /api/search/weights

  • Read

    GET /api/segment

    Fetch *all* `Segments`.

    GET /api/segment

  • Read

    GET /api/segment/{id}

    Fetch `Segment` with ID.

    GET /api/segment/{id}

  • Read

    GET /api/session/illustration/{key}

    Fetch the uploaded image of a custom illustration setting, e.g. `login-page-illustration-custom`.

    GET /api/session/illustration/{key}

  • Read

    GET /api/session/password_reset_token_valid

    Check if a password reset token is valid and isn't expired.

    GET /api/session/password_reset_token_valid

  • Read

    GET /api/session/properties

    Get all properties and their values. These are the specific `Settings` that are readable by the current user, or are public if no user is logged in.

    GET /api/session/properties

  • Read

    GET /api/setting

    Get all `Settings` and their values. You must be a superuser or have `setting` permission to do this. For non-superusers, a list of visible settings and values can be retrieved using the /api/session/properties endpoint.

    GET /api/setting

  • Read

    GET /api/setting/{key}

    Fetch a single `Setting`.

    GET /api/setting/{key}

  • Read

    GET /api/slack/app-info

    Returns the Slack app_id and team_id. Used by the frontend to construct direct links to the Slack app settings page.

    GET /api/slack/app-info

  • Read

    GET /api/slack/manifest

    Returns the JSON manifest file that should be used to bootstrap new Slack apps

    GET /api/slack/manifest

  • Read

    GET /api/table

    Get all `Tables`. Optional filters: - `can-query=true` - filter to only tables the user can execute queries against - `include-editable-data-model=true` - filter to only tables the user can edit the data model of, instead of the tables the user can read

    GET /api/table

  • Read

    GET /api/table/card__:id/fks

    Return FK info for the 'virtual' table for a Card. This is always empty, so this endpoint serves mainly as a placeholder to avoid having to change anything on the frontend.

    GET /api/table/card__:id/fks

  • Read

    GET /api/table/card__:id/query_metadata

    Return metadata for the 'virtual' table for a Card.

    GET /api/table/card__:id/query_metadata

  • Read

    GET /api/table/{id}

    Get `Table` with ID.

    GET /api/table/{id}

  • Read

    GET /api/table/{id}/fks

    Get all foreign keys whose destination is a `Field` that belongs to this `Table`.

    GET /api/table/{id}/fks

  • Read

    GET /api/table/{id}/query_metadata

    Get metadata about a `Table` useful for running queries. Returns DB, fields, field FKs, and field values. Passing `include_hidden_fields=true` will include any hidden `Fields` in the response. Defaults to `false`. Passing `include_sensitive_fields=true` will include any sensitive `Fields` in the response if the current user has data model editing permissions for the table. Defaults to `false`. Passing `include_editable_data_model=true` will check that the current user has data model editing permissions for the table, even when Remote Sync makes the table read-only, while `false` checks that they can read the table. Defaults to `false`. The response includes `can_write`, which is false when the user cannot edit the table's metadata right now. These options are provided for use in the data model pages.

    GET /api/table/{id}/query_metadata

  • Read

    GET /api/table/{table-id}/data

    Get the data for the given table

    GET /api/table/{table-id}/data

  • Read

    GET /api/task

    Fetch a list of recent tasks stored as Task History

    GET /api/task

  • Read

    GET /api/task/info

    Return raw data about all scheduled tasks (i.e., Quartz Jobs and Triggers).

    GET /api/task/info

  • Read

    GET /api/task/runs

    List task runs with optional filters. Returns runs with hydrated entity names and task counts.

    GET /api/task/runs

  • Read

    GET /api/task/runs/entities

    Get distinct entities that have task runs for a given run type. Used for populating entity filter picker.

    GET /api/task/runs/entities

  • Read

    GET /api/task/runs/{id}

    Get a single task run with all its child tasks.

    GET /api/task/runs/{id}

  • Read

    GET /api/task/unique-tasks

    Returns possibly empty vector of unique task names in alphabetical order. It is expected that number of unique tasks is small, hence no need for pagination. If that changes this endpoint and function that powers it should reflect that.

    GET /api/task/unique-tasks

  • Read

    GET /api/task/{id}

    Get `TaskHistory` entry with ID.

    GET /api/task/{id}

  • Read

    GET /api/tiles/{card-id}/{zoom}/{x}/{y}

    Generates a single tile image for a saved Card.

    GET /api/tiles/{card-id}/{zoom}/{x}/{y}

  • Read

    GET /api/tiles/{dashboard-id}/dashcard/{dashcard-id}/card/{card-id}/{zoom}/{x}/{y}

    Generates a single tile image for a dashcard.

    GET /api/tiles/{dashboard-id}/dashcard/{dashcard-id}/card/{card-id}/{zoom}/{x}/{y}

  • Read

    GET /api/tiles/{zoom}/{x}/{y}

    Generates a single tile image for an ad-hoc query.

    GET /api/tiles/{zoom}/{x}/{y}

  • Read

    GET /api/timeline

    Fetch a list of `Timeline`s. Can include `archived=true` to return archived timelines.

    GET /api/timeline

  • Read

    GET /api/timeline-event/{id}

    Fetch the [[TimelineEvent]] with `id`.

    GET /api/timeline-event/{id}

  • Read

    GET /api/timeline/collection/root

    Fetch the root Collection's timelines.

    GET /api/timeline/collection/root

  • Read

    GET /api/timeline/collection/{id}

    Fetch a specific Collection's timelines.

    GET /api/timeline/collection/{id}

  • Read

    GET /api/timeline/{id}

    Fetch the `Timeline` with `id`. Include `include=events` to unarchived events included on the timeline. Add `archived=true` to return all events on the timeline, both archived and unarchived.

    GET /api/timeline/{id}

  • Read

    GET /api/transform

    Get a list of transforms.

    GET /api/transform

  • Read

    GET /api/transform-dag-run/{run-id}/transform-runs

    Get the transform runs that made up a specific DAG run.

    GET /api/transform-dag-run/{run-id}/transform-runs

  • Read

    GET /api/transform-job

    Get all transform jobs.

    GET /api/transform-job

  • Read

    GET /api/transform-job/{job-id}

    Get a transform job by ID.

    GET /api/transform-job/{job-id}

  • Read

    GET /api/transform-job/{job-id}/runs

    Get paginated run history for a transform job.

    GET /api/transform-job/{job-id}/runs

  • Read

    GET /api/transform-job/{job-id}/runs/{run-id}/transform-runs

    Get the transform runs that made up a specific job run.

    GET /api/transform-job/{job-id}/runs/{run-id}/transform-runs

  • Read

    GET /api/transform-job/{job-id}/transforms

    Get the transforms of job specified by the job's ID.

    GET /api/transform-job/{job-id}/transforms

  • Read

    GET /api/transform-tag

    Get a list of all transform tags.

    GET /api/transform-tag

  • Read

    GET /api/transform/run

    Get transform runs based on a set of filter params.

    GET /api/transform/run

  • Read

    GET /api/transform/run/{run-id}

    Get a transform run by ID.

    GET /api/transform/run/{run-id}

  • Read

    GET /api/transform/runs

    Paginated unified run history: every row is a root run — a job run, a manual DAG-reprocess run, or a standalone transform run — never a member run of a job/DAG (those are listed by `GET /run` and the per-run `transform-runs` endpoints). `types` selects which kinds to include (all by default); `transform-ids` narrows to runs that ran any of the given transforms. The remaining filters work as in `GET /run`.

    GET /api/transform/runs

  • Read

    GET /api/transform/{id}

    Get a specific transform.

    GET /api/transform/{id}

  • Read

    GET /api/transform/{id}/dag-transforms

    Preview the transforms a DAG reprocess from this transform would run (see `POST /:id/run-dag`), in execution order.

    GET /api/transform/{id}/dag-transforms

  • Read

    GET /api/transform/{id}/dependencies

    Get the dependencies of a specific transform.

    GET /api/transform/{id}/dependencies

  • Read

    GET /api/typed-schemas/v1/typescript

    Generate a TypeScript semantic schema module.

    GET /api/typed-schemas/v1/typescript

  • Read

    GET /api/user

    Fetch a list of `Users` for admins or group managers. By default returns only active users for admins/data-analysts and only active users within groups that the group manager is managing for group managers. - If `status` is `deactivated`, include deactivated users only. - If `status` is `all`, include all users (active and inactive). - Also supports `include_deactivated`, which if true, is equivalent to `status=all`; If is false, is equivalent to `status=active`. `status` and `include_deactivated` requires superuser permissions. - `include_deactivated` is a legacy alias for `status` and will be removed in a future release, users are advised to use `status` for better support and flexibility. If both params are passed, `status` takes precedence. - if a `tenant_id` is passed, only users with that tenant_id will be returned. For users with segmented permissions, return only themselves. Takes `limit`, `offset` for pagination. Takes `query` for filtering on first name, last name, email. Also takes `group_id`, which filters on group id. If the user is a sandboxed user, only return themselves regardless of the query parameters.

    GET /api/user

  • Read

    GET /api/user-key-value/namespace/{namespace}

    Returns all KV pairs in a given namespace for the current user

    GET /api/user-key-value/namespace/{namespace}

  • Read

    GET /api/user-key-value/namespace/{namespace}/key/{key}

    Get a value for the user

    GET /api/user-key-value/namespace/{namespace}/key/{key}

  • Read

    GET /api/user/current

    Fetch the current `User`.

    GET /api/user/current

  • Read

    GET /api/user/recipients

    Fetch a list of `Users`. Returns only active users. Meant for non-admins unlike GET /api/user. - If user-visibility is :all or the user is an admin, include all users. - If user-visibility is :group, include only users in the same group (excluding the all users group). - If user-visibility is :none or the user is sandboxed, include only themselves.

    GET /api/user/recipients

  • Read

    GET /api/user/{id}

    Fetch a `User`. You must be fetching yourself *or* be a superuser *or* a Group Manager. Only personal users can be fetched this way; API-key users and the internal user 404.

    GET /api/user/{id}

  • Read

    GET /api/util/random_token

    Return a cryptographically secure random 32-byte token, encoded as a hexadecimal string. Intended for use when creating a value for `embedding-secret-key`.

    GET /api/util/random_token

  • Write

    POST /api/action

    Create a new action.

    POST /api/action

  • Write

    POST /api/action/{action-id}/execute/values

    Fetches the values for filling in execution parameters. Pass PK parameters and values to select. Parameters are sent in the request body rather than the query string so their values stay out of URLs and logs.

    POST /api/action/{action-id}/execute/values

  • Write

    PUT /api/action/{id}

    Update an Action.

    PUT /api/action/{id}

  • Write

    POST /api/action/{id}/execute

    Execute the Action. `parameters` should be the mapped dashboard parameters with values.

    POST /api/action/{id}/execute

  • Write

    POST /api/action/{id}/public_link

    Generate publicly-accessible links for this Action. Returns UUID to be used in public links. (If this Action has already been shared, it will return the existing public link rather than creating a new one.) Public sharing must be enabled.

    POST /api/action/{id}/public_link

  • Write

    POST /api/activity/recents

    Adds a model to the list of recently selected items.

    POST /api/activity/recents

  • Write

    POST /api/agent/v1/collection

    Create a new Collection. Pass `parent_collection_id` to nest under another collection; omit for a root-level collection. The caller must have write access to the parent (or root, if no parent given).

    POST /api/agent/v1/collection

  • Write

    POST /api/agent/v1/construct-native-query

    Construct a native (raw SQL) query against a database. Wraps `sql` into a serialized native query and returns it base64-encoded — the same `{"query": <base64>}` shape as /v2/construct-query — so it can be saved as a question via /v1/question (the MCP layer swaps the base64 for a `query_handle`). This endpoint does NOT execute the SQL; to run native SQL ad hoc use /v1/execute-sql. The MBQL execution endpoints (/v1/execute, /v2/query) reject native queries by design, so a handle produced here is for saving, not for /v2/query execution.

    POST /api/agent/v1/construct-native-query

  • Write

    POST /api/agent/v1/dashboard

    Create a new dashboard, optionally populated with saved questions. Pass `question_ids` to add existing saved questions as cards on the dashboard. Cards are automatically positioned on the grid based on their display type. If `collection_id` is omitted the dashboard is saved to the caller's personal collection. Pass an explicit `null` to save it to the root collection. The response `collection_path` is the saved location.

    POST /api/agent/v1/dashboard

  • Write

    POST /api/agent/v1/execute

    Execute an MBQL query and return results. Accepts a base64-encoded MBQL query (as returned by /v1/construct-query) and executes it, returning results with column metadata. Response format: - On success: {:data {:cols [...] :rows [...]} :row_count N :status :completed :running_time M} - On failure: {:status :failed :error "message" ...} Standard userspace query limits are enforced (2000 rows for simple queries, 10000 for aggregated).

    POST /api/agent/v1/execute

  • Write

    POST /api/agent/v1/execute-sql

    Execute a raw SQL query against a database. Returns rows + column metadata. Requires the user to have native-query permission on the target database; the QP middleware enforces this. The instance-level `mcp-execute-sql-enabled` setting must also be on (it is by default).

    POST /api/agent/v1/execute-sql

  • Write

    POST /api/agent/v1/metric

    Save a previously constructed query as a named metric. A metric is a reusable aggregation (a `Card` of type `metric`): the underlying query must have exactly one aggregation and at most one date/datetime grouping. The `query` parameter accepts a `query_handle` (UUID) returned by `construct_query`, or a base64-encoded MBQL string. MCP callers should always use the handle. Optionally specify display type, description, collection, and visualization settings. If `collection_id` is omitted the metric is saved to the caller's personal collection. Pass an explicit `null` to save it to the root collection. The response `collection_path` is the saved location.

    POST /api/agent/v1/metric

  • Write

    POST /api/agent/v1/question

    Save a previously constructed query as a named question (card). The `query` parameter accepts a `query_handle` (UUID) returned by `construct_query` (MBQL) or `construct_native_query` (native SQL), or a base64-encoded query string. MCP callers should always use the handle. Optionally specify display type, description, collection, and visualization settings. If `collection_id` is omitted the question is saved to the caller's personal collection. Pass an explicit `null` to save it to the root collection. The response `collection_path` is the saved location. Saving a native (raw SQL) query requires native-query permission on the target database.

    POST /api/agent/v1/question

  • Write

    POST /api/agent/v1/question/{id}/query

    Run a saved question (card) and return its results. Executes the query stored on the card under the caller's permissions and returns rows + column metadata — the same response shape as /v1/execute. Parameterized questions are NOT supported: if the card declares parameters or input template tags (field filters / variables), this returns a 400. Run it in Metabase, or save a parameterless version.

    POST /api/agent/v1/question/{id}/query

  • Write

    POST /api/agent/v1/read-resource

    Read one or more Metabase resources via metabase:// URI patterns. Dispatches into the shared URI resolver in `metabase.metabot.tools.resources`, which validates URIs, fetches entities with per-URI permission checks, and returns a map of `{:resources ... :output ...}`. Up to 5 URIs per call.

    POST /api/agent/v1/read-resource

  • Write

    POST /api/agent/v1/search

    Search for tables, models, metrics, saved questions, dashboards, and collections. Supports both term-based and semantic search queries. Results are ranked using Reciprocal Rank Fusion when both query types are provided.

    POST /api/agent/v1/search

  • Write

    POST /api/agent/v2/construct-query

    Construct an MBQL query from a portable MBQL 5 representations JSON payload. The body is `{"query": <external-query>}` where `<external-query>` is a JSON object matching `::lib.schema/external-query` — see the `construct_notebook_query` tool documentation for the format reference. Returns a base64-encoded MBQL query that can be executed via /v1/execute or paginated via /v2/query.

    POST /api/agent/v2/construct-query

  • Write

    POST /api/agent/v2/query

    Execute a portable MBQL 5 representations JSON query and stream the results, with continuation-token pagination. Accepts either a JSON body (same shape as /v2/construct-query) or a `continuation_token` from a previous response. Returns results with column metadata and an optional `continuation_token` for fetching the next page.

    POST /api/agent/v2/query

  • Write

    POST /api/analytics-proxy

    Forward a Snowplow `tp2` payload to `sp.metabase.com` server-side, bypassing the customer page's `connect-src` CSP.

    POST /api/analytics-proxy

  • Write

    POST /api/analytics/internal

    Receive a batch of internal analytics events from the frontend and record them as Prometheus metrics. Only the counters in [[frontend-metrics]] can be incremented; other events are ignored.

    POST /api/analytics/internal

  • Write

    POST /api/api-key

    Create a new API key (and an associated `User`) with the provided name and group ID.

    POST /api/api-key

  • Write

    PUT /api/api-key/{id}

    Update an API key by changing its group and/or its name

    PUT /api/api-key/{id}

  • Write

    PUT /api/api-key/{id}/regenerate

    Regenerate an API Key

    PUT /api/api-key/{id}/regenerate

  • Write

    POST /api/apps/{slug}/draft

    Create or reuse a data app draft before its first repository import.

    POST /api/apps/{slug}/draft

  • Write

    POST /api/apps/{slug}/query

    Resolve an authored data-app query definition into a serializable Metabase query.

    POST /api/apps/{slug}/query

  • Write

    POST /api/apps/{slug}/query-table-dependencies

    Return the tables read by already-saved queries, including ones reached only through an implicit join. Sync copies models, actions and metrics whose queries it never resolves through `/query`, and only this metadata-based lookup sees an implicit join: the id of a table reached through a foreign key appears nowhere in the query itself.

    POST /api/apps/{slug}/query-table-dependencies

  • Write

    PUT /api/apps/{slug}/table-dependencies

    Store the tables used by the resources from a successful data app resource synchronization.

    PUT /api/apps/{slug}/table-dependencies

  • Write

    POST /api/apps/{slug}/user-permission-warnings

    Return warnings for users who cannot access every table used by a data app.

    POST /api/apps/{slug}/user-permission-warnings

  • Write

    PUT /api/bookmark/ordering

    Sets the order of bookmarks for user.

    PUT /api/bookmark/ordering

  • Write

    POST /api/bookmark/{model}/{id}

    Create a new bookmark for user.

    POST /api/bookmark/{model}/{id}

  • Write

    PUT /api/cache

    Store cache configuration.

    PUT /api/cache

  • Write

    POST /api/card

    Create a new `Card`. Card `type` can be `question`, `metric`, or `model`.

    POST /api/card

  • Write

    POST /api/card/pivot/{card-id}/query

    Run the query associated with a Card.

    POST /api/card/pivot/{card-id}/query

  • Write

    POST /api/card/{card-id}/public_link

    Generate publicly-accessible links for this Card. Returns UUID to be used in public links. (If this Card has already been shared, it will return the existing public link rather than creating a new one.) Public sharing must be enabled.

    POST /api/card/{card-id}/public_link

  • Write

    POST /api/card/{card-id}/query

    Run the query associated with a Card. When `stored_result_id` is supplied, serve the cached snapshot instead of re-running the query and optionally re-sorts the rows via the `sort` body param.

    POST /api/card/{card-id}/query

  • Write

    POST /api/card/{card-id}/query/{export-format}

    Run the query associated with a Card, and return its results as a file in the specified format. `csv_include_bom`, `parameters`, `pivot-results?` and `format-rows?` should be passed as application/x-www-form-urlencoded form content or json in the body. This is because this endpoint is normally used to power 'Download Results' buttons that use HTML `form` actions).

    POST /api/card/{card-id}/query/{export-format}

  • Write

    PUT /api/card/{id}

    Update a `Card`.

    PUT /api/card/{id}

  • Write

    POST /api/card/{id}/copy

    Copy a `Card`, with the new name 'Copy of _name_'

    POST /api/card/{id}/copy

  • Write

    POST /api/cards/dashboards

    Get the dashboards that multiple cards appear in. The response is a sequence of maps, each of which has a `card_id` and `dashboards`. `dashboard` may include an `:error` key, either `:unreadable-dashboard` or `:unwritable-dashboard`. In the case of an `unreadable-dashboard` the dashboard details (name, ID) will NOT be present.

    POST /api/cards/dashboards

  • Write

    POST /api/cards/move

    Moves a number of Cards to a single collection or dashboard. For now, just either succeed or fail as a batch - we can think more about error handling later down the road.

    POST /api/cards/move

  • Write

    POST /api/channel

    Create a channel

    POST /api/channel

  • Write

    POST /api/channel/test

    Test a channel connection

    POST /api/channel/test

  • Write

    PUT /api/channel/{id}

    Update a channel

    PUT /api/channel/{id}

  • Write

    POST /api/cloud-migration

    Initiate a new cloud migration.

    POST /api/cloud-migration

  • Write

    POST /api/collection

    Create a new Collection.

    POST /api/collection

  • Write

    POST /api/collection/root/move-dashboard-question-candidates

    Move candidate cards to the dashboards they appear in (for the root collection)

    POST /api/collection/root/move-dashboard-question-candidates

  • Write

    POST /api/collection/{id}/move-dashboard-question-candidates

    Move candidate cards to the dashboards they appear in.

    POST /api/collection/{id}/move-dashboard-question-candidates

  • Write

    POST /api/comment

    Create a new comment

    POST /api/comment

  • Write

    PUT /api/comment/{comment-id}

    Update a comment

    PUT /api/comment/{comment-id}

  • Write

    POST /api/comment/{comment-id}/reaction

    Toggle a reaction on a comment

    POST /api/comment/{comment-id}/reaction

  • Write

    POST /api/dashboard

    Create a new Dashboard.

    POST /api/dashboard

  • Write

    POST /api/dashboard/pivot/{dashboard-id}/dashcard/{dashcard-id}/card/{card-id}/query

    Run a pivot table query for a specific DashCard.

    POST /api/dashboard/pivot/{dashboard-id}/dashcard/{dashcard-id}/card/{card-id}/query

  • Write

    POST /api/dashboard/save

    Save a denormalized description of dashboard.

    POST /api/dashboard/save

  • Write

    POST /api/dashboard/save/collection/{parent-collection-id}

    Save a denormalized description of dashboard into collection with ID `:parent-collection-id`.

    POST /api/dashboard/save/collection/{parent-collection-id}

  • Write

    POST /api/dashboard/{dashboard-id}/dashcard/{dashcard-id}/card/{card-id}/query

    Run the query associated with a Saved Question (`Card`) in the context of a `Dashboard` that includes it.

    POST /api/dashboard/{dashboard-id}/dashcard/{dashcard-id}/card/{card-id}/query

  • Write

    POST /api/dashboard/{dashboard-id}/dashcard/{dashcard-id}/card/{card-id}/query/{export-format}

    Run the query associated with a Saved Question (`Card`) in the context of a `Dashboard` that includes it, and return its results as a file in the specified format. `parameters` should be passed as query parameter encoded as a serialized JSON string (this is because this endpoint is normally used to power 'Download Results' buttons that use HTML `form` actions).

    POST /api/dashboard/{dashboard-id}/dashcard/{dashcard-id}/card/{card-id}/query/{export-format}

  • Write

    POST /api/dashboard/{dashboard-id}/dashcard/{dashcard-id}/execute

    Execute the associated Action in the context of a `Dashboard` and `DashboardCard` that includes it. `parameters` should be the mapped dashboard parameters with values. `extra_parameters` should be the extra, user entered parameter values.

    POST /api/dashboard/{dashboard-id}/dashcard/{dashcard-id}/execute

  • Write

    POST /api/dashboard/{dashboard-id}/dashcard/{dashcard-id}/execute/values

    Fetches the values for filling in execution parameters. Pass PK parameters and values to select. Parameters are sent in the request body rather than the query string so their values stay out of URLs and logs.

    POST /api/dashboard/{dashboard-id}/dashcard/{dashcard-id}/execute/values

  • Write

    POST /api/dashboard/{dashboard-id}/public_link

    Generate publicly-accessible links for this Dashboard. Returns UUID to be used in public links. (If this Dashboard has already been shared, it will return the existing public link rather than creating a new one.) Public sharing must be enabled.

    POST /api/dashboard/{dashboard-id}/public_link

  • Write

    POST /api/dashboard/{from-dashboard-id}/copy

    Copy a Dashboard.

    POST /api/dashboard/{from-dashboard-id}/copy

  • Write

    PUT /api/dashboard/{id}

    Update a Dashboard, and optionally the `dashcards` and `tabs` of a Dashboard. The request body should be a JSON object with the same structure as the response from `GET /api/dashboard/:id`.

    PUT /api/dashboard/{id}

  • Write

    PUT /api/dashboard/{id}/cards

    (DEPRECATED -- Use the `PUT /api/dashboard/:id` endpoint instead.) Update `Cards` and `Tabs` on a Dashboard. Request body should have the form: {:cards [{:id ... ; DashboardCard ID :size_x ... :size_y ... :row ... :col ... :parameter_mappings ... :series [{:id 123 ...}]} ...] :tabs [{:id ... ; DashboardTab ID :name ...}]}

    PUT /api/dashboard/{id}/cards

  • Write

    POST /api/dashboard/{id}/pdf

    Render Dashboard with ID to a PDF (server-side, the same way dashboard subscriptions render charts) and stream it back as a file download. `parameters` are runtime parameter overrides -- the same shape the dashcard query endpoints accept: a sequence of `{:id ... :value ...}` maps, or a JSON-encoded string of that sequence (the string form lets an HTML `<form>` action drive the download). Parameters left unspecified fall back to the dashboard's own defaults. `paper_size` is `"a4"` (default) or `"letter"`.

    POST /api/dashboard/{id}/pdf

  • Write

    POST /api/data-studio/table/edit

    Bulk updating tables.

    POST /api/data-studio/table/edit

  • Write

    POST /api/data-studio/table/rescan-values

    Batch version of /table/:id/rescan_values. Takes an abstract table selection as /table/edit does.

    POST /api/data-studio/table/rescan-values

  • Write

    POST /api/data-studio/table/selection

    Gets information about selected tables

    POST /api/data-studio/table/selection

  • Write

    POST /api/data-studio/table/sync-schema

    Batch version of /table/:id/sync_schema. Takes an abstract table selection as /table/edit does. - Currently checks policy before returning (so you might receive a 4xx on e.g. AuthZ policy failure) - The sync itself is however, asynchronous. This call may return before all tables synced.

    POST /api/data-studio/table/sync-schema

  • Write

    POST /api/database

    Add a new `Database`.

    POST /api/database

  • Write

    POST /api/database/sample_database

    Add the sample database as a new `Database`.

    POST /api/database/sample_database

  • Write

    POST /api/database/validate

    Validate that we can connect to a database given a set of details.

    POST /api/database/validate

  • Write

    PUT /api/database/{id}

    Update a `Database`.

    PUT /api/database/{id}

  • Write

    POST /api/database/{id}/dismiss_spinner

    Manually set the initial sync status of the `Database` and corresponding tables to be `complete` (see #20863)

    POST /api/database/{id}/dismiss_spinner

  • Write

    POST /api/database/{id}/rescan_values

    Trigger a manual scan of the field values for this `Database`.

    POST /api/database/{id}/rescan_values

  • Write

    POST /api/database/{id}/sync_schema

    Trigger a manual update of the schema metadata for this `Database`.

    POST /api/database/{id}/sync_schema

  • Write

    POST /api/dataset

    Execute a query and retrieve the results in the usual format. The query will not use the cache.

    POST /api/dataset

  • Write

    POST /api/dataset/native

    Fetch a native version of an MBQL query.

    POST /api/dataset/native

  • Write

    POST /api/dataset/parameter/remapping

    Return the remapped parameter values for cards or dashboards that are being edited.

    POST /api/dataset/parameter/remapping

  • Write

    POST /api/dataset/parameter/search/{query}

    Return parameter values for cards or dashboards that are being edited. Expects a query string at `?query=foo`.

    POST /api/dataset/parameter/search/{query}

  • Write

    POST /api/dataset/parameter/values

    Return parameter values for cards or dashboards that are being edited.

    POST /api/dataset/parameter/values

  • Write

    POST /api/dataset/pivot

    Generate a pivoted dataset for an ad-hoc query

    POST /api/dataset/pivot

  • Write

    POST /api/dataset/query_metadata

    Get all of the required query metadata for an ad-hoc query. You can pass `{:settings {:include-sensitive-fields true}}` in the query to include fields with visibility_type :sensitive in the response.

    POST /api/dataset/query_metadata

  • Write

    POST /api/dataset/{export-format}

    Execute a query and download the result data as a file in the specified format.

    POST /api/dataset/{export-format}

  • Write

    POST /api/document

    Create a new `Document`.

    POST /api/document

  • Write

    PUT /api/document/{document-id}

    Updates an existing `Document`.

    PUT /api/document/{document-id}

  • Write

    POST /api/document/{document-id}/card/{card-id}/query/{export-format}

    Download query results for a Card embedded in a Document. Returns query results in the requested format. The user must have read access to the document to download results. If the card's query fails, standard query error responses are returned. Route parameters: - document-id: ID of the document containing the card - card-id: ID of the card to download results from - export-format: Output format (csv, xlsx, json) Body parameters (snake_case): - parameters: Optional query parameters (array of maps or JSON string) - format_rows: Whether to apply formatting to results (boolean, default false) - pivot_results: Whether to pivot results (boolean, default false)

    POST /api/document/{document-id}/card/{card-id}/query/{export-format}

  • Write

    POST /api/document/{document-id}/public-link

    Generate a publicly-accessible UUID for a Document. Creates a public link that allows viewing the Document without authentication. If the Document already has a public UUID, returns the existing one rather than generating a new one. This enables sharing the Document via `GET /api/ee/public/document/:uuid`. Returns a map containing `:uuid` (the public UUID string). Requires superuser permissions. Public sharing must be enabled via the `enable-public-sharing` setting.

    POST /api/document/{document-id}/public-link

  • Write

    POST /api/document/{from-document-id}/copy

    Copy a Document.

    POST /api/document/{from-document-id}/copy

  • Write

    POST /api/ee/action-v2/execute

    Execute an action with a single input. Takes: - `action` - an identifier or an expression for what we want to execute. - `scope` - where the action is being invoked from. - `input` - a single map. currently these are typically a database table row pk, or query result. - `params` (optional) - a map of values for the parameters taken by the action's mapping. The `input` and `params` are used by the relevant mapping to calculate a map argument to the underlying action fn. If there is no mapping, `params` are simply used as overrides for `input`. Returns the outputs from the performed action.

    POST /api/ee/action-v2/execute

  • Write

    POST /api/ee/action-v2/execute-bulk

    Execute an action with multiple inputs. This is typically more efficient than calling execute with each input individually, for example by performing batch SQL operations. Takes: - `action` - an identifier or an expression for what we want to execute. - `scope` - where the action is being invoked from. - `inputs` - a list of maps. currently these are typically a database table row pk, or query result. - `params` (optional) - a map of values for the parameters taken by the action's mapping. The `inputs` and `params` are used by the relevant mapping to calculate a list of args for the underlying action fn. If there is no mapping, `params` are simply used as overrides for each map within `inputs`. Returns the outputs from the performed action.

    POST /api/ee/action-v2/execute-bulk

  • Write

    POST /api/ee/action-v2/execute-form

    Temporary endpoint for describing an actions parameters such that they can be presented correctly in a modal ahead of execution.

    POST /api/ee/action-v2/execute-form

  • Write

    PUT /api/ee/advanced-permissions/application/graph

    Do a batch update of Application Permissions by passing a modified graph.

    PUT /api/ee/advanced-permissions/application/graph

  • Write

    PUT /api/ee/ai-controls/permissions

    Update metabot permissions for all groups. Upserts each permission entry and returns the full permissions list with defaults filled in.

    PUT /api/ee/ai-controls/permissions

  • Write

    POST /api/ee/audit-app/analytics-dev/export

    Export analytics content as a .tar.gz file for local development. Only available when MB_ANALYTICS_DEV_MODE=true. Returns a tarball containing the analytics YAMLs in canonical format, ready to commit to source control. Requires superuser permissions.

    POST /api/ee/audit-app/analytics-dev/export

  • Write

    POST /api/ee/cloud-add-ons/{product-type}

    Purchase an add-on.

    POST /api/ee/cloud-add-ons/{product-type}

  • Write

    POST /api/ee/cloud-proxy/{operation-id}

    Proxy a call to the Metabase Store API via harbormaster client. This endpoint is used only for hosted instances, and calls Harbormaster Store using a OpenAPI client. :operation-id is the operation-id of the Harbormaster Store endpoint. All parameters for the operation are taken in the POST body.

    POST /api/ee/cloud-proxy/{operation-id}

  • Write

    POST /api/ee/content-translation/upload-dictionary

    Upload a CSV of content translations

    POST /api/ee/content-translation/upload-dictionary

  • Write

    POST /api/ee/custom-viz-plugin

    Register a new custom visualization plugin from an uploaded tar.gz bundle. The archive must contain `metabase-plugin.json` at the root and `dist/index.js` for the JS bundle. The only static asset served from the bundle is the manifest `icon` (under `dist/assets/`); plugins do not ship arbitrary assets. The plugin's `identifier` is taken from the manifest's `name` field.

    POST /api/ee/custom-viz-plugin

  • Write

    POST /api/ee/custom-viz-plugin/dev

    Register a dev-only custom visualization plugin from a local dev server. No bundle upload is required — files are served from the dev server URL. Requires custom viz plugin dev mode to be enabled.

    POST /api/ee/custom-viz-plugin/dev

  • Write

    PUT /api/ee/custom-viz-plugin/{id}

    Update a custom visualization plugin. Currently only `enabled` may be toggled.

    PUT /api/ee/custom-viz-plugin/{id}

  • Write

    POST /api/ee/custom-viz-plugin/{id}/refresh

    Re-fetch the manifest from the dev server for a dev-only plugin. For uploaded plugins this is a no-op — to update an upload-backed plugin, PUT a new bundle to `/:id/bundle`.

    POST /api/ee/custom-viz-plugin/{id}/refresh

  • Write

    POST /api/ee/data-studio/table/publish-tables

    Set collection for each of selected tables and all upstream dependencies recursively.

    POST /api/ee/data-studio/table/publish-tables

  • Write

    POST /api/ee/data-studio/table/unpublish-tables

    Unset collection for each of selected tables and all downstream dependents recursively.

    POST /api/ee/data-studio/table/unpublish-tables

  • Write

    POST /api/ee/database-replication/connection/{database-id}

    Create a new PG replication connection for the specified database.

    POST /api/ee/database-replication/connection/{database-id}

  • Write

    POST /api/ee/database-replication/connection/{database-id}/preview

    Return info about pg-replication connection that is about to be created.

    POST /api/ee/database-replication/connection/{database-id}/preview

  • Write

    POST /api/ee/database-routing/destination-database

    Create new Destination Databases. Note that unlike the normal `POST /api/database` endpoint, this endpoint does not test that the database is actually reachable before adding it — destination details are not required to be valid at creation time, and an unreachable destination is fine.

    POST /api/ee/database-routing/destination-database

  • Write

    PUT /api/ee/database-routing/router-database/{id}

    Updates an existing Database with the `user_attribute` to route on. Will either: - turn an existing Database into a Router database - change the `user_attribute` used to route for an existing Router database, or - turn a Router database into a regular Database depending on the value of `user_attribute`

    PUT /api/ee/database-routing/router-database/{id}

  • Write

    PUT /api/ee/email/override

    Update multiple cloud email Settings. You must be a superuser or have `setting` permission to do this. Calling this automatically sets `cloud-smtp-enabled` to true if the settings are valid.

    PUT /api/ee/email/override

  • Write

    POST /api/ee/gsheets/connection

    Hook up a new google drive folder or sheet that will be watched and have its content ETL'd into Metabase.

    POST /api/ee/gsheets/connection

  • Write

    POST /api/ee/gsheets/connection/sync

    Force a sync of the connection now. Returns the gsheets shape, with the attached datawarehouse db id at `:db_id`.

    POST /api/ee/gsheets/connection/sync

  • Write

    POST /api/ee/library

    Creates the Library if it doesn't exist. Returns the created collection. Requires data analyst or superuser permissions.

    POST /api/ee/library

  • Write

    POST /api/ee/mfa/enroll

    Start TOTP enrollment for the current user. Requires the account password (LDAP users re-bind against the directory) and the `:multi-factor-auth` feature. Returns the Base32 `secret` and an `otpauth_uri` for QR display; enrollment is not active until confirmed with a live code.

    POST /api/ee/mfa/enroll

  • Write

    POST /api/ee/mfa/enroll/confirm

    Confirm TOTP enrollment by verifying a code from the authenticator app. Activates the second factor and returns the single-use recovery codes — the only time they exist in plaintext.

    POST /api/ee/mfa/enroll/confirm

  • Write

    POST /api/ee/remote-sync/create-branch

    Create a new branch from the current remote-sync branch and switches the current remote-sync branch to it. Requires superuser permissions.

    POST /api/ee/remote-sync/create-branch

  • Write

    POST /api/ee/remote-sync/export

    Export the current state of the Remote Sync collection to a Source. This endpoint will: - Fetch the latest changes from the source - Create a branch or subdirectory (depending on source support) If no branch is supplied, use the configured export branch - Export the Remote Sync collection via serialization to the branch or subdirectory - Commit the changes if possible - Sync to the source if possible Requires superuser permissions.

    POST /api/ee/remote-sync/export

  • Write

    POST /api/ee/remote-sync/import

    Import Metabase content from configured Remote Sync source. This endpoint will: - Fetch the latest changes from the configured source - Load the updated content using the serialization/deserialization system If `force=false` (default) and there are unsaved changes in the Remote Sync collection, the import returns a 400 response. Requires superuser permissions.

    POST /api/ee/remote-sync/import

  • Write

    PUT /api/ee/remote-sync/settings

    Update Remote Sync related settings. You must be a superuser to do this.

    PUT /api/ee/remote-sync/settings

  • Write

    POST /api/ee/remote-sync/stash

    Stashes changes to a new branch, and changes the current branch to it. Requires superuser permissions.

    POST /api/ee/remote-sync/stash

  • Write

    POST /api/ee/remote-sync/test-connection

    Test whether the Remote Sync credentials can reach the git repository. When called with an empty body, validates the currently saved URL and token. When the body provides `remote-sync-url` or `remote-sync-token`, those override the saved values — useful for verifying a new Personal Access Token before saving. An obfuscated token (matching the existing token's masked representation) is treated as "unchanged" and the stored token value is used for the test. Only validates connection and authentication; branch existence is not checked here. Returns `{:status :success}` on success. On failure, returns a 400 with a user-friendly error message describing the connection problem. Requires superuser permissions.

    POST /api/ee/remote-sync/test-connection

  • Write

    POST /api/ee/scim/api_key

    Create a new SCIM API key, or refresh one that already exists. When called for the first time, this is equivalent to enabling SCIM.

    POST /api/ee/scim/api_key

  • Write

    POST /api/ee/scim/v2/Groups

    Create a single group, and populates it if necessary.

    POST /api/ee/scim/v2/Groups

  • Write

    PUT /api/ee/scim/v2/Groups/{id}

    Update a group.

    PUT /api/ee/scim/v2/Groups/{id}

  • Write

    POST /api/ee/scim/v2/Users

    Create a single user.

    POST /api/ee/scim/v2/Users

  • Write

    PUT /api/ee/scim/v2/Users/{id}

    Update a user.

    PUT /api/ee/scim/v2/Users/{id}

  • Write

    POST /api/ee/security-center/acknowledge

    Acknowledge multiple security advisories. Skips already-acknowledged advisories.

    POST /api/ee/security-center/acknowledge

  • Write

    POST /api/ee/security-center/sync

    Trigger an async advisory sync + re-evaluation. Returns immediately. If a sync is already running, the request is a no-op.

    POST /api/ee/security-center/sync

  • Write

    POST /api/ee/security-center/test-notification

    Send a test notification through the given Security Center channels. The request body lets callers pass the unsaved notification config from the dialog so the test reflects current form state, not the persisted settings. Both fields are optional; when omitted, the saved setting is used.

    POST /api/ee/security-center/test-notification

  • Write

    POST /api/ee/serialization/export

    Serialize and retrieve Metabase instance. Outputs `.tar.gz` file with serialization results and an `export.log` file. On error outputs serialization logs directly.

    POST /api/ee/serialization/export

  • Write

    POST /api/ee/serialization/import

    Deserialize Metabase instance from an archive generated by /export. Parameters: - `file`: archive encoded as `multipart/form-data` (required). Returns logs of deserialization.

    POST /api/ee/serialization/import

  • Write

    POST /api/ee/support-access-grant

    Create a new support access grant. Requires superuser permissions. Only one active grant can exist at a time.

    POST /api/ee/support-access-grant

  • Write

    POST /api/ee/tenant

    Create a new Tenant

    POST /api/ee/tenant

  • Write

    PUT /api/ee/tenant/{id}

    Update a tenant, can set name, attributes, or whether this tenant is active.

    PUT /api/ee/tenant/{id}

  • Write

    PUT /api/ee/transforms-python/library/{path}

    Update the Python library source code for user modules.

    PUT /api/ee/transforms-python/library/{path}

  • Write

    POST /api/ee/transforms-python/test-run

    Evaluate an ad-hoc python transform on a sample of input data. Intended for short runs for early feedback. Input/output/timeout limits apply.

    POST /api/ee/transforms-python/test-run

  • Write

    POST /api/ee/transforms/{id}/inspect/{lens-id}/query

    Execute a query in the context of a transform inspector lens.

    POST /api/ee/transforms/{id}/inspect/{lens-id}/query

  • Write

    POST /api/eid-translation/translate

    Translate entity IDs to model IDs.

    POST /api/eid-translation/translate

  • Write

    PUT /api/email

    Update multiple email Settings. You must be a superuser or have `setting` permission to do this.

    PUT /api/email

  • Write

    POST /api/email/test

    Send a test email using the SMTP Settings. You must be a superuser or have `setting` permission to do this. Returns `{:ok true}` if we were able to send the message successfully, otherwise a standard 400 error response.

    POST /api/email/test

  • Write

    POST /api/embed-mcp/drills

    Stash a base64-encoded MBQL query for the iframe's pending drill-through and return a handle UUID the iframe will thread into the agent message so the `render_drill_through` tool can fetch it.

    POST /api/embed-mcp/drills

  • Write

    POST /api/embed-mcp/feedback

    Persist MCP Apps visualization feedback.

    POST /api/embed-mcp/feedback

  • Write

    POST /api/embed-theme

    Create a new embedding theme.

    POST /api/embed-theme

  • Write

    POST /api/embed-theme/seed-defaults

    Seed default embedding themes on first call, using the payloads built by the frontend from the `METABASE_LIGHT_THEME` / `METABASE_DARK_THEME` constants. Idempotent: guarded by the `default-embedding-themes-seeded` setting. Once flipped, subsequent calls are no-ops even if the admin has since deleted the seeded themes, so deletions are preserved.

    POST /api/embed-theme/seed-defaults

  • Write

    PUT /api/embed-theme/{id}

    Update an embedding theme.

    PUT /api/embed-theme/{id}

  • Write

    POST /api/embed-theme/{id}/copy

    Copy an embedding theme.

    POST /api/embed-theme/{id}/copy

  • Write

    POST /api/field/table-ids

    Get unique Table IDs for a list of Field IDs.

    POST /api/field/table-ids

  • Write

    PUT /api/field/{id}

    Update `Field` with ID.

    PUT /api/field/{id}

  • Write

    POST /api/field/{id}/dimension

    Sets the dimension for the given field at ID

    POST /api/field/{id}/dimension

  • Write

    POST /api/field/{id}/rescan_values

    Manually trigger an update for the FieldValues for this Field. Only applies to Fields that are eligible for FieldValues.

    POST /api/field/{id}/rescan_values

  • Write

    POST /api/field/{id}/values

    Update the fields values and human-readable values for a `Field` whose semantic type is `category`/`city`/`state`/`country` or whose base type is `type/Boolean`. The human-readable values are optional.

    POST /api/field/{id}/values

  • Write

    POST /api/frontend-errors

    Endpoint for the frontend to report errors. Increments a Prometheus counter with the given `type` label.

    POST /api/frontend-errors

  • Write

    POST /api/glossary

    Create a new glossary entry.

    POST /api/glossary

  • Write

    PUT /api/glossary/{id}

    Update an existing glossary entry.

    PUT /api/glossary/{id}

  • Write

    PUT /api/google/settings

    Update Google Sign-In related settings. You must be a superuser to do this.

    PUT /api/google/settings

  • Write

    POST /api/index/request

    Create an index request on a transform's target table. Requires write access to the transform. Listed via `GET /`, not under `/request`.

    POST /api/index/request

  • Write

    PUT /api/ldap/settings

    Update LDAP related settings. You must be a superuser to do this.

    PUT /api/ldap/settings

  • Write

    POST /api/llm/extract-sources

    Parse native query sources and return referenced tables and cards/models. Uses Macaw to parse the SQL, resolves table names to IDs, and returns permission-filtered tables with column metadata. Card and model references are extracted from native query template tags. This is a lightweight endpoint that does not trigger fingerprinting or field value fetching.

    POST /api/llm/extract-sources

  • Write

    POST /api/llm/generate-sql

    Generate SQL from a natural language prompt. Requires: - LLM to be configured (Anthropic API key set in admin settings) - At least one table reference (explicit @mention or implicit from source_sql) - A database_id parameter Returns generated SQL and the list of tables used for context.

    POST /api/llm/generate-sql

  • Write

    POST /api/llm/providers

    Create a provider connection. The credentials are verified before the connection is saved.

    POST /api/llm/providers

  • Write

    PUT /api/llm/providers/{key}

    Update a provider connection. Secret fields the client echoes back masked keep their stored value.

    PUT /api/llm/providers/{key}

  • Write

    POST /api/logger/adjustment

    Temporarily adjust the log levels.

    POST /api/logger/adjustment

  • Write

    POST /api/measure

    Create a new `Measure`. The Measure's table is derived from its `definition`.

    POST /api/measure

  • Write

    PUT /api/measure/{id}

    Update a `Measure` with ID.

    PUT /api/measure/{id}

  • Write

    POST /api/metabot/agent-streaming

    Send a chat message to the LLM via the AI Proxy.

    POST /api/metabot/agent-streaming

  • Write

    POST /api/metabot/conversations/{id}/fork

    Fork a conversation at an assistant message, returning a brand-new conversation that copies the thread from the start up to and including that message. Only the conversation's originator may fork it, for now. `message_id` is the `external_id` of the message to fork at, which must be a live, finished, non-errored assistant message. The clone gets fresh conversation/message ids, fresh timestamps, the current user as owner, and zeroed token usage so the fork does not double-count tokens in the analytics views.

    POST /api/metabot/conversations/{id}/fork

  • Write

    POST /api/metabot/conversations/{id}/saved-entity

    Save a Metabot-generated chart from this conversation as a card, stamping the card's origin columns in the same request — used by the inline chart's manual Save button, which runs outside any agent turn. Creating and stamping together (rather than stamping after a separate `POST /api/card`) means the card and its origin cannot desync when the follow-up request is lost. Accessible to any participant in the conversation or to any superuser.

    POST /api/metabot/conversations/{id}/saved-entity

  • Write

    POST /api/metabot/document/generate-content

    Create a new piece of content to insert into the document. Kept for backwards compatibility; now uses the native Clojure agent.

    POST /api/metabot/document/generate-content

  • Write

    POST /api/metabot/feedback

    Persist Metabot feedback.

    POST /api/metabot/feedback

  • Write

    PUT /api/metabot/metabot/{id}

    Update a metabot instance

    PUT /api/metabot/metabot/{id}

  • Write

    POST /api/metabot/slack/events

    Respond to activities in Slack

    POST /api/metabot/slack/events

  • Write

    POST /api/metabot/slack/interactive

    Handle interactive payloads from Slack (button clicks, modal submissions).

    POST /api/metabot/slack/interactive

  • Write

    POST /api/metabot/source-feedback

    Persist Metabot source feedback.

    POST /api/metabot/source-feedback

  • Write

    POST /api/metric/breakout-values

    Fetch distinct breakout dimension values for a metric or measure definition. Accepts the same definition format as POST /dataset. Returns extracted values and column metadata for the breakout dimension.

    POST /api/metric/breakout-values

  • Write

    POST /api/metric/dataset

    Execute a metric or measure-based query and stream the results. Request body requires a `definition` object containing: - expression: A metric math expression tree (leaf reference or arithmetic) Examples: [:metric {:lib/uuid "a"} 42], [:- {} [:metric {:lib/uuid "a"} 1] [:metric {:lib/uuid "b"} 2]] - filters (optional): Per-instance filters keyed by :lib/uuid from the expression - projections (optional): Typed projections keyed by source type and ID

    POST /api/metric/dataset

  • Write

    POST /api/metric/{id}/dimension/add

    Add dimensions to a metric. Each entry carries its UUID and mapping target; the column-derived fields and mapping are recomputed server-side. Returns the updated list of dimensions.

    POST /api/metric/{id}/dimension/add

  • Write

    POST /api/metric/{id}/dimension/reorder

    Persist a new ordering for a metric's dimensions. `dimension_ids` is the desired order; dimensions not listed keep their relative order after the listed ones. Returns the updated list of added dimensions.

    POST /api/metric/{id}/dimension/reorder

  • Write

    POST /api/metric/{id}/dimension/{dimension-key}

    Update a metric dimension's `display_name`, `description`, `default_temporal_unit`, and/or source column. `source` is a `{type, field-id}`; the new column must have the same effective type.

    POST /api/metric/{id}/dimension/{dimension-key}

  • Write

    POST /api/model-index

    Create ModelIndex.

    POST /api/model-index

  • Write

    POST /api/moderation-review

    Create a new `ModerationReview`.

    POST /api/moderation-review

  • Write

    POST /api/mt/gtap

    Create a new GTAP.

    POST /api/mt/gtap

  • Write

    POST /api/mt/gtap/validate

    Validate a sandbox which may not have yet been saved. This runs the same validation that is performed when the sandbox is saved, but doesn't actually save the sandbox.

    POST /api/mt/gtap/validate

  • Write

    PUT /api/mt/user/{id}/attributes

    Update the `login_attributes` for a User. Only personal users can have attributes.

    PUT /api/mt/user/{id}/attributes

  • Write

    POST /api/native-query-snippet

    Create a new `NativeQuerySnippet`.

    POST /api/native-query-snippet

  • Write

    PUT /api/native-query-snippet/{id}

    Update an existing `NativeQuerySnippet`.

    PUT /api/native-query-snippet/{id}

  • Write

    POST /api/notification

    Create a new notification, return the created notification.

    POST /api/notification

  • Write

    POST /api/notification/send

    Send an unsaved notification.

    POST /api/notification/send

  • Write

    POST /api/notification/unsubscribe/undo

    Allow non-users to undo an unsubscribe from notifications, with the hash given through email.

    POST /api/notification/unsubscribe/undo

  • Write

    PUT /api/notification/{id}

    Update a notification, can also update its subscriptions, handlers. Return the updated notification. `creator_id` (owner) can be reassigned here only by superusers (e.g. the admin 'Edit alert' modal's owner picker). `mi/can-update?` rejects a non-superuser reassignment attempt with 403; the model's `before-update` hook is the backstop. Echoing back the unchanged value is fine.

    PUT /api/notification/{id}

  • Write

    POST /api/notification/{id}/send

    Send a notification by id.

    POST /api/notification/{id}/send

  • Write

    POST /api/notify/db/attached_datawarehouse

    Sync the attached datawarehouse. Can provide in the body: - table_name and schema_name: both strings. Will look for an existing table and sync it, otherwise will try to find a new table with that name and sync it. If it cannot find a table it will throw an error. If table_name is empty or blank, will sync the entire database. - synchronous?: is a boolean value to indicate if this should block on the result.

    POST /api/notify/db/attached_datawarehouse

  • Write

    POST /api/notify/db/{id}

    Notification about a potential schema change to one of our `Databases`. Caller can optionally specify a `:table_id` or `:table_name` in the body to limit updates to a single `Table`. Optional Parameter `:scan` can be `"full"` or `"schema"` for a full sync or a schema sync, available regardless if a `:table_id` or `:table_name` is passed. This endpoint is secured by an API key that needs to be passed as a `X-METABASE-APIKEY` header which needs to be defined in the `MB_API_KEY` [environment variable](https://www.metabase.com/docs/latest/configuring-metabase/environment-variables.html#mb_api_key)

    POST /api/notify/db/{id}

  • Write

    POST /api/notify/db/{id}/new-table

    Sync a new table without running a full database sync. Requires `schema_name` and `table_name`. Will throw an error if the table already exists in Metabase or cannot be found.

    POST /api/notify/db/{id}/new-table

  • Write

    POST /api/osi/ai-context/reconcile

    Force a reconcile of the library entity index against the application database, blocking until a reconcile covering this call finishes. Returns the index mutations (`index`) separately from execution timing (`execution`: how long the run waited to start, then how long it ran). This call never reuses a reconcile already in progress (it may have started before your latest change); it starts one if the index is idle, otherwise it queues a single follow-up that any other waiting calls share. Requires the library entity-retrieval feature; returns a 400 when the index is unavailable (the feature isn't licensed, or the pgvector store or embedding backend isn't configured).

    POST /api/osi/ai-context/reconcile

  • Write

    PUT /api/permissions/graph

    Do a batch update of Permissions by passing in a modified graph. This should return the same graph, in the same format, that you got from `GET /api/permissions/graph`, with any changes made in the wherever necessary. This modified graph must correspond to the `PermissionsGraph` schema. If successful, this endpoint returns the updated permissions graph; use this as a base for any further modifications. Revisions to the permissions graph are tracked. If you fetch the permissions graph and some other third-party modifies it before you can submit you revisions, the endpoint will instead make no changes and return a 409 (Conflict) response. In this case, you should fetch the updated graph and make desired changes to that. The optional `sandboxes` key contains a list of sandboxes that should be created or modified in conjunction with this permissions graph update. Since data sandboxing is an Enterprise Edition-only feature, a 402 (Payment Required) response will be returned if this key is present and the server is not running the Enterprise Edition, and/or the `:sandboxes` feature flag is not present. If the skip-graph query param is truthy, then the graph will not be returned.

    PUT /api/permissions/graph

  • Write

    POST /api/permissions/group

    Create a new `PermissionsGroup`.

    POST /api/permissions/group

  • Write

    PUT /api/permissions/group/{group-id}

    Update the name of a `PermissionsGroup`.

    PUT /api/permissions/group/{group-id}

  • Write

    POST /api/permissions/membership

    Add a `User` to a `PermissionsGroup`. Returns updated list of members belonging to the group.

    POST /api/permissions/membership

  • Write

    PUT /api/permissions/membership/{id}

    Update a Permission Group membership. Returns the updated record.

    PUT /api/permissions/membership/{id}

  • Write

    POST /api/persist/card/{card-id}/persist

    Mark the model (card) as persisted. Runs the query and saves it to the database backing the card and hot swaps this query in place of the model's query.

    POST /api/persist/card/{card-id}/persist

  • Write

    POST /api/persist/card/{card-id}/refresh

    Refresh the persisted model caching `card-id`.

    POST /api/persist/card/{card-id}/refresh

  • Write

    POST /api/persist/database/{id}/persist

    Attempt to enable model persistence for a database. If already enabled returns a generic 204.

    POST /api/persist/database/{id}/persist

  • Write

    POST /api/persist/enable

    Enable global setting to allow databases to persist models.

    POST /api/persist/enable

  • Write

    POST /api/persist/set-refresh-schedule

    Set the cron schedule to refresh persisted models. Shape should be JSON like {cron: "0 30 1/8 * * ? *"}.

    POST /api/persist/set-refresh-schedule

  • Write

    POST /api/product-feedback

    Endpoint to provide feedback from the product

    POST /api/product-feedback

  • Write

    POST /api/public/action/{uuid}/execute

    Execute the Action. `parameters` should be the mapped dashboard parameters with values.

    POST /api/public/action/{uuid}/execute

  • Write

    POST /api/public/dashboard/{uuid}/dashcard/{dashcard-id}/card/{card-id}/{export-format}

    Fetch the results of running a publicly-accessible Card belonging to a Dashboard and return the data in one of the export formats. Does not require auth credentials. Public sharing must be enabled.

    POST /api/public/dashboard/{uuid}/dashcard/{dashcard-id}/card/{card-id}/{export-format}

  • Write

    POST /api/public/dashboard/{uuid}/dashcard/{dashcard-id}/execute

    Execute the associated Action in the context of a `Dashboard` and `DashboardCard` that includes it. `parameters` should be the mapped dashboard parameters with values.

    POST /api/public/dashboard/{uuid}/dashcard/{dashcard-id}/execute

  • Write

    POST /api/public/dashboard/{uuid}/dashcard/{dashcard-id}/execute/values

    Fetches the values for filling in execution parameters. Pass PK parameters and values to select. Parameters are sent in the request body rather than the query string so their values stay out of URLs and logs.

    POST /api/public/dashboard/{uuid}/dashcard/{dashcard-id}/execute/values

  • Write

    POST /api/public/document/{uuid}/card/{card-id}/{export-format}

    Fetch a Card embedded in a public Document and return query results in the specified format. Does not require auth credentials. Public sharing must be enabled.

    POST /api/public/document/{uuid}/card/{card-id}/{export-format}

  • Write

    POST /api/pulse

    Create a new `Pulse`.

    POST /api/pulse

  • Write

    POST /api/pulse/test

    Test send an unsaved pulse.

    POST /api/pulse/test

  • Write

    POST /api/pulse/unsubscribe/undo

    Allow non-users to undo an unsubscribe from pulses/subscriptions, with the hash given through email.

    POST /api/pulse/unsubscribe/undo

  • Write

    PUT /api/pulse/{id}

    Update a Pulse with `id`.

    PUT /api/pulse/{id}

  • Write

    POST /api/search/force-reindex

    This will trigger an immediate reindexing, if we are using search index.

    POST /api/search/force-reindex

  • Write

    POST /api/search/re-init

    This will blow away any search indexes, re-create, and re-populate them.

    POST /api/search/re-init

  • Write

    PUT /api/search/weights

    Update the current weights being used to rank the search results

    PUT /api/search/weights

  • Write

    POST /api/segment

    Create a new `Segment`. The Segment's table is derived from its `definition`.

    POST /api/segment

  • Write

    PUT /api/segment/{id}

    Update a `Segment` with ID.

    PUT /api/segment/{id}

  • Write

    POST /api/session

    Login.

    POST /api/session

  • Write

    POST /api/session/google_auth

    Login with Google Auth.

    POST /api/session/google_auth

  • Write

    POST /api/session/mfa/enroll

    Complete a two-step login by *enrolling* a second factor for the first time. This happens when a user without MFA enrolled logs in for the first time after the instance starts *requiring* MFA. After confirming their email and password, the user will see an MFA enrollment screen, served by `POST /session` after it accepts the password. That body includes a single-use, signed JWT similar to the one used by an MFA challenge (see `POST /mfa/verify` above) as well as the details needed to show the QR code for enrolling MFA. This request must include the JWT (`enrollment_token`) and the OTP itself (`code`). If enrollment succeeds, both the pending enrollment and the JWT's `jti` are consumed, preventing reuse of either. A successful response **sets the session cookie, logging the user in!** Its body contains the recovery codes in the same form as enrollment when already authenticated.

    POST /api/session/mfa/enroll

  • Write

    POST /api/session/mfa/send-email-otp

    Email a one-time code as a fallback second factor (for a user who lost their authenticator but still has recovery codes disabled or unavailable). Requires a valid challenge token from `POST /api/session`; the code is single-use with a 10-minute expiry and is accepted by `POST /mfa/verify` like any other code.

    POST /api/session/mfa/send-email-otp

  • Write

    POST /api/session/mfa/verify

    Complete a two-step login by verifying a one-time code. Takes the `challenge_token` returned by `POST /api/session` and either the 6-digit `code` from the user's authenticator app, one of their single-use recovery codes, or an emailed one-time code; on success sets the session cookie.

    POST /api/session/mfa/verify

  • Write

    POST /api/session/password-check

    Endpoint that checks if the supplied password meets the currently configured password complexity rules.

    POST /api/session/password-check

  • Write

    PUT /api/setting

    Update multiple `Settings` values. If called by a non-superuser, only user-local settings can be updated.

    PUT /api/setting

  • Write

    POST /api/setup

    Special endpoint for creating the first user during setup. This endpoint both creates the user AND logs them in and returns a session ID. This endpoint can also be used to set specific settings from the setup flow.

    POST /api/setup

  • Write

    POST /api/slack/bug-report

    Send diagnostic information to the configured Slack channels. Requires bug reporting to be enabled. The report is attributed to the current user when `diagnosticInfo.reporter` is true, and anonymous otherwise. The `{name, email}` form of `reporter` that clients before 0.64 send is treated as true; the identity in it is ignored.

    POST /api/slack/bug-report

  • Write

    PUT /api/table

    Update all `Table` in `ids`. Deprecated, should use PUT /table/edit from now on.

    PUT /api/table

  • Write

    PUT /api/table/{id}

    Update `Table` with ID.

    PUT /api/table/{id}

  • Write

    POST /api/table/{id}/append-csv

    Inserts the rows of an uploaded CSV file into the table identified by `:id`. The table must have been created by uploading a CSV file. The file may be at most 50 MB; larger uploads are rejected with a 413 response.

    POST /api/table/{id}/append-csv

  • Write

    PUT /api/table/{id}/fields/order

    Reorder fields

    PUT /api/table/{id}/fields/order

  • Write

    POST /api/table/{id}/rescan_values

    Manually trigger an update for the FieldValues for the Fields belonging to this Table. Only applies to Fields that are eligible for FieldValues.

    POST /api/table/{id}/rescan_values

  • Write

    POST /api/table/{id}/sync_schema

    Trigger a manual update of the schema metadata for this `Table`.

    POST /api/table/{id}/sync_schema

  • Write

    POST /api/timeline

    Create a new [[Timeline]].

    POST /api/timeline

  • Write

    POST /api/timeline-event

    Create a new [[TimelineEvent]].

    POST /api/timeline-event

  • Write

    PUT /api/timeline-event/{id}

    Update a [[TimelineEvent]].

    PUT /api/timeline-event/{id}

  • Write

    POST /api/transform

    Create a new transform.

    POST /api/transform

  • Write

    POST /api/transform-job

    Create a new transform job.

    POST /api/transform-job

  • Write

    PUT /api/transform-job/active

    Activate or deactivate every transform job. Inactive jobs do not run on schedule. Manual runs via `POST /api/transform-job/:job-id/run` ignore this flag. Reports per-job outcome counts: `:updated` (successfully flipped) and `:failed` (raised an error during the flip — the row update or Quartz write failed and was logged).

    PUT /api/transform-job/active

  • Write

    PUT /api/transform-job/{job-id}

    Update a transform job.

    PUT /api/transform-job/{job-id}

  • Write

    POST /api/transform-job/{job-id}/run

    Run a transform job manually. Returns a 202 with the created `job_run_id`, or a nil `job_run_id` when nothing was run (the job is already running, or has no transforms). By default, fresh pulled-in dependencies are skipped; pass `run_all` to force-refresh the whole plan.

    POST /api/transform-job/{job-id}/run

  • Write

    POST /api/transform-tag

    Create a new transform tag.

    POST /api/transform-tag

  • Write

    PUT /api/transform-tag/{tag-id}

    Update a transform tag.

    PUT /api/transform-tag/{tag-id}

  • Write

    PUT /api/transform/{id}

    Update a transform.

    PUT /api/transform/{id}

  • Write

    POST /api/transform/{id}/run

    Run a transform.

    POST /api/transform/{id}/run

  • Write

    POST /api/transform/{id}/run-dag

    Trigger a DAG-reprocess run starting from a single transform: runs the transform and every transform in its transitive dependency closure. Returns a 202 with the created `dag_run_id`, or a nil `dag_run_id` when nothing was run (a DAG run for this transform is already in progress, or the closure is empty). `direction` selects which transforms are included: - `upstream` — the seed transform plus all transforms it depends on - `downstream` — the seed transform plus all transforms that depend on it

    POST /api/transform/{id}/run-dag

  • Write

    POST /api/upload/csv

    Create a table and model populated with the values from the attached CSV. Returns the model ID if successful. The file may be at most 50 MB; larger uploads are rejected with a 413 response.

    POST /api/upload/csv

  • Write

    POST /api/user

    Create a new `User`, return a 400 if the email address is already taken

    POST /api/user

  • Write

    PUT /api/user-key-value/namespace/{namespace}/key/{key}

    Upsert a KV-pair for the user

    PUT /api/user-key-value/namespace/{namespace}/key/{key}

  • Write

    PUT /api/user/{id}/modal/{modal}

    Indicate that a user has been informed about the vast intricacies of 'the' Query Builder.

    PUT /api/user/{id}/modal/{modal}

  • Write

    PUT /api/user/{id}/password

    Update a user's password.

    PUT /api/user/{id}/password

  • Write

    PUT /api/user/{id}/reactivate

    Reactivate user at `:id`

    PUT /api/user/{id}/reactivate

  • Destructive

    DELETE /api/action/{action-id}

    Delete an Action.

    DELETE /api/action/{action-id}

  • Destructive

    PUT /api/agent/v1/dashboard/{id}

    Update a dashboard. Patch semantics - only fields you pass are changed. Metadata: `name`, `description`, `collection_id`, `archived`. Archiving is a soft delete - there is no delete endpoint. It can be reversed by setting `archived: false`. Dashcard mutations are submitted under `dashcards` as a list of `{action: add|add_heading|add_text|update_text|remove|move, ...}` entries applied in order. `add` requires `card_id`; `add_heading` and `add_text` require `text` (Markdown for text cards); `update_text`, `remove`, and `move` require `dashcard_id`.

    PUT /api/agent/v1/dashboard/{id}

  • Destructive

    PUT /api/agent/v1/metric/{id}

    Update a saved metric. Patch semantics - only fields that you pass are changed. Set `collection_id` to move the metric to a different collection. Set `archived: true` to archive. Archiving is a soft delete - there is no delete endpoint. It can be reversed by setting `archived: false`. Pass `query` (a query_handle from construct_query, or a base64 MBQL string) to replace the underlying query; the replacement must still be a valid metric (exactly one aggregation, at most one date/datetime grouping). The target card must be a metric — use `update_question` for questions.

    PUT /api/agent/v1/metric/{id}

  • Destructive

    PUT /api/agent/v1/question/{id}

    Update a saved question (card). Patch semantics - only fields that you pass are changed. Set `collection_id` to move the card to a different collection. Set `archived: true` to archive. Archiving is a soft delete - there is no delete endpoint. It can be reversed by setting `archived: false`. Pass `query` (a query_handle from construct_query or construct_native_query, or a base64 query string) to replace the underlying query. Replacing it with a native (raw SQL) query requires native-query permission on the target database.

    PUT /api/agent/v1/question/{id}

  • Destructive

    DELETE /api/alert/{id}/subscription

    For users to unsubscribe themselves from the given alert.

    DELETE /api/alert/{id}/subscription

  • Destructive

    DELETE /api/api-key/{id}

    Delete an ApiKey

    DELETE /api/api-key/{id}

  • Destructive

    PUT /api/apps/{slug}

    Enable or disable a single data app. Disabled apps are not served.

    PUT /api/apps/{slug}

  • Destructive

    DELETE /api/apps/{slug}

    Remove a single data app (its row and cached bundle). Intended for clearing out apps left behind by a repository that is no longer connected: while a repo is connected, an app still in it is re-materialized by the next sync, and one no longer in it is pruned by that sync anyway.

    DELETE /api/apps/{slug}

  • Destructive

    DELETE /api/bookmark/{model}/{id}

    Delete a bookmark. Will delete a bookmark assigned to the user making the request by model and id.

    DELETE /api/bookmark/{model}/{id}

  • Destructive

    DELETE /api/cache

    Delete cache configurations.

    DELETE /api/cache

  • Destructive

    POST /api/cache/invalidate

    Invalidate cache entries. Use it like `/api/cache/invalidate?database=1&dashboard=15` (any number of database/dashboard/question can be supplied). `&include=overrides` controls whenever you want to invalidate cache for a specific cache configuration without touching all nested configurations, or you want your invalidation to trickle down to every card.

    POST /api/cache/invalidate

  • Destructive

    POST /api/card/collections

    Bulk update endpoint for Card Collections. Move a set of `Cards` with `card_ids` into a `Collection` with `collection_id`, or remove them from any Collections by passing a `null` `collection_id`.

    POST /api/card/collections

  • Destructive

    DELETE /api/card/{id}

    Hard delete a Card. To soft delete, use `PUT /api/queries/:id`

    DELETE /api/card/{id}

  • Destructive

    PUT /api/cloud-migration/cancel

    Cancel any ongoing cloud migrations, if any.

    PUT /api/cloud-migration/cancel

  • Destructive

    PUT /api/collection/graph

    Do a batch update of Collections Permissions by passing in a modified graph. Will overwrite parts of the graph that are present in the request, and leave the rest unchanged. If the `force` query parameter is `true`, a `revision` number is not required. The provided graph will be persisted as-is, and has the potential to clobber other writes that happened since the last read. If the `skip_graph` query parameter is `true`, it will only return the current revision, not the entire permissions graph.

    PUT /api/collection/graph

  • Destructive

    PUT /api/collection/{id}

    Modify an existing Collection, including archiving or unarchiving it, or moving it.

    PUT /api/collection/{id}

  • Destructive

    DELETE /api/collection/{id}

    Deletes a collection permanently

    DELETE /api/collection/{id}

  • Destructive

    DELETE /api/comment/{comment-id}

    Soft delete a comment

    DELETE /api/comment/{comment-id}

  • Destructive

    DELETE /api/dashboard/{id}

    Hard delete a Dashboard. To soft delete, use `PUT /api/dashboard/:id` This will remove also any questions/models/segments/metrics that use this database.

    DELETE /api/dashboard/{id}

  • Destructive

    POST /api/data-studio/table/discard-values

    Batch version of /table/:id/discard_values. Takes an abstract table selection as /table/edit does.

    POST /api/data-studio/table/discard-values

  • Destructive

    DELETE /api/database/{id}

    Delete a `Database`.

    DELETE /api/database/{id}

  • Destructive

    POST /api/database/{id}/discard_values

    Discards all saved field values for this `Database`.

    POST /api/database/{id}/discard_values

  • Destructive

    DELETE /api/document/{document-id}

    Permanently deletes an archived Document.

    DELETE /api/document/{document-id}

  • Destructive

    DELETE /api/ee/advanced-permissions/impersonation/{id}

    Delete a Connection Impersonation entry.

    DELETE /api/ee/advanced-permissions/impersonation/{id}

  • Destructive

    POST /api/ee/ai-controls/permissions/advanced

    Switch to group-level permissions. Removes the permissions of All Users and All tenant users, so nobody has access until they are in a group that grants it.

    POST /api/ee/ai-controls/permissions/advanced

  • Destructive

    DELETE /api/ee/ai-controls/permissions/advanced

    Switch back to simple permissions. Removes the permissions of every group other than Administrators, All Users and All tenant users.

    DELETE /api/ee/ai-controls/permissions/advanced

  • Destructive

    PUT /api/ee/ai-controls/usage/group/{group-id}

    Set or update the metabot usage limit for a specific group. Pass `max_usage: null` to remove the limit.

    PUT /api/ee/ai-controls/usage/group/{group-id}

  • Destructive

    PUT /api/ee/ai-controls/usage/instance

    Set or update the instance-wide metabot usage limit. Pass `max_usage: null` to remove the limit (unlimited).

    PUT /api/ee/ai-controls/usage/instance

  • Destructive

    PUT /api/ee/ai-controls/usage/tenant/{tenant-id}

    Set or update the metabot usage limit for a specific tenant. Pass `max_usage: null` to remove the limit.

    PUT /api/ee/ai-controls/usage/tenant/{tenant-id}

  • Destructive

    DELETE /api/ee/audit-app/user/{id}/subscriptions

    Delete all Alert and DashboardSubscription subscriptions for a User (i.e., so they will no longer receive them). Archive all Alerts and DashboardSubscriptions created by the User. Only allowed for admins or for the current user.

    DELETE /api/ee/audit-app/user/{id}/subscriptions

  • Destructive

    DELETE /api/ee/cloud-add-ons/{product-type}

    Remove an add-on.

    DELETE /api/ee/cloud-add-ons/{product-type}

  • Destructive

    DELETE /api/ee/custom-viz-plugin/{id}

    Remove a custom visualization plugin and evict its on-disk cache.

    DELETE /api/ee/custom-viz-plugin/{id}

  • Destructive

    PUT /api/ee/custom-viz-plugin/{id}/bundle

    Replace the bundle for an existing plugin. Accepts a multipart tar.gz upload in the same format as the `POST /` endpoint. The manifest's `name` field must match the plugin's existing `identifier`.

    PUT /api/ee/custom-viz-plugin/{id}/bundle

  • Destructive

    PUT /api/ee/custom-viz-plugin/{id}/dev-url

    Set or clear the dev base URL for a plugin (e.g. `http://localhost:5174`). The bundle is fetched from `{base}/index.js` and assets from `{base}/assets/{name}`. Persisted to the database so it survives server restarts. Requires custom viz plugin dev mode to be enabled.

    PUT /api/ee/custom-viz-plugin/{id}/dev-url

  • Destructive

    DELETE /api/ee/database-replication/connection/{database-id}

    Delete PG replication connection for the specified database.

    DELETE /api/ee/database-replication/connection/{database-id}

  • Destructive

    DELETE /api/ee/email/override

    Clear all cloud email related settings. You must be a superuser or have `setting` permission to do this.

    DELETE /api/ee/email/override

  • Destructive

    DELETE /api/ee/gsheets/connection

    Disconnect the google service account. There is only one (or zero) at the time of writing.

    DELETE /api/ee/gsheets/connection

  • Destructive

    POST /api/ee/mfa/admin/remove

    Admin: remove *another* user's two-factor enrollment entirely — the lockout escape hatch for a lost authenticator with no recovery codes. Never feature-gated (a lapsed license must not make lockouts permanent). The affected user is notified by email. They re-enroll from scratch — there is nothing to "reset", the secret lives on their device. Removing your *own* enrollment is deliberately refused here: this is a user-management endpoint that takes no second factor (the target can't produce one — that's why an admin is removing it), and the admin UI never collects a code. Self-removal goes through the normal Security page, which re-auths with a fresh factor via `/disable`. Without this guard a hijacked admin session could strip its own 2FA with only a cookie, turning transient access into a permanent password bypass.

    POST /api/ee/mfa/admin/remove

  • Destructive

    POST /api/ee/mfa/disable

    Disable two-factor authentication for the current user. Re-auth is a fresh second factor — a TOTP code or an unused recovery code — never just the password.

    POST /api/ee/mfa/disable

  • Destructive

    POST /api/ee/mfa/recovery-codes

    Regenerate the current user's recovery codes, invalidating the entire previous set. Re-auth is a fresh second factor — a TOTP code or an unused recovery code — so a stolen password alone can never rotate the codes. The plaintext codes are returned exactly once; only hashes are stored.

    POST /api/ee/mfa/recovery-codes

  • Destructive

    POST /api/ee/remote-sync/current-task/cancel

    Cancels the current task if one is running

    POST /api/ee/remote-sync/current-task/cancel

  • Destructive

    POST /api/ee/replacement/check-replace-source

    Check whether a source entity can be replaced by a target entity. Returns compatibility errors describing column mismatches, type mismatches, primary key mismatches, and foreign key mismatches.

    POST /api/ee/replacement/check-replace-source

  • Destructive

    POST /api/ee/replacement/replace-model-with-transform

    Create a transform from a model, execute it, and replace all usages of the model with the output table. Un-persists the model and converts it to a saved question. Returns 202 with a run_id for polling. If there is an error during the transform execution, no replacement will be performed and the model will remain unchanged. If there is an error during the source swap, the transform and the output table will be retained, and the model will remain unchanged. We cannot delete the transform or the output table because they can be used by other queries at this point.

    POST /api/ee/replacement/replace-model-with-transform

  • Destructive

    POST /api/ee/replacement/replace-source

    Replace all usages of a source entity with a target entity asynchronously. Returns 202 with a run_id for polling. Returns 409 if a replacement is already running.

    POST /api/ee/replacement/replace-source

  • Destructive

    POST /api/ee/replacement/runs/{id}/cancel

    Cancel a running source replacement.

    POST /api/ee/replacement/runs/{id}/cancel

  • Destructive

    DELETE /api/ee/scim/v2/Groups/{id}

    Delete a group.

    DELETE /api/ee/scim/v2/Groups/{id}

  • Destructive

    PATCH /api/ee/scim/v2/Users/{id}

    Activate or deactivate a user. Supports specific replace operations, but not arbitrary patches.

    PATCH /api/ee/scim/v2/Users/{id}

  • Destructive

    POST /api/ee/security-center/{advisory-id}/acknowledge

    Acknowledge a security advisory. Stops repeat notifications.

    POST /api/ee/security-center/{advisory-id}/acknowledge

  • Destructive

    PUT /api/ee/support-access-grant/{id}/revoke

    Revoke an existing support access grant. Requires superuser permissions. Any admin can revoke any grant.

    PUT /api/ee/support-access-grant/{id}/revoke

  • Destructive

    DELETE /api/ee/upload-management/tables/{id}

    Delete the uploaded table from the database, optionally archiving cards for which it is the primary source.

    DELETE /api/ee/upload-management/tables/{id}

  • Destructive

    DELETE /api/email

    Clear all email related settings. You must be a superuser or have `setting` permission to do this.

    DELETE /api/email

  • Destructive

    DELETE /api/embed-theme/{id}

    Delete an embedding theme.

    DELETE /api/embed-theme/{id}

  • Destructive

    DELETE /api/field/{id}/dimension

    Remove the dimension associated to field at ID

    DELETE /api/field/{id}/dimension

  • Destructive

    POST /api/field/{id}/discard_values

    Discard the FieldValues belonging to this Field. Only applies to fields that have FieldValues. If this Field's Database is set up to automatically sync FieldValues, they will be recreated during the next cycle.

    POST /api/field/{id}/discard_values

  • Destructive

    DELETE /api/glossary/{id}

    Delete a glossary entry.

    DELETE /api/glossary/{id}

  • Destructive

    PUT /api/index/request/{id}

    Replace the structured definition of an index request, marking it update-pending.

    PUT /api/index/request/{id}

  • Destructive

    DELETE /api/index/request/{id}

    Mark an index request `delete-pending`. The physical index is only dropped when the target table is next rebuilt (pending changes force a full run), so the row stays visible in this state until that rebuild removes it.

    DELETE /api/index/request/{id}

  • Destructive

    DELETE /api/llm/providers/{key}

    Delete a provider connection.

    DELETE /api/llm/providers/{key}

  • Destructive

    DELETE /api/logger/adjustment

    Undo any log level adjustments.

    DELETE /api/logger/adjustment

  • Destructive

    DELETE /api/metabot/metabot/{id}/prompt-suggestions

    Delete all prompt suggestions for the metabot instance with `id`.

    DELETE /api/metabot/metabot/{id}/prompt-suggestions

  • Destructive

    POST /api/metabot/metabot/{id}/prompt-suggestions/regenerate

    Remove any existing prompt suggestions for the Metabot with `id` and generate new ones. The response `:status` is `:generated` (with a `:prompt_count`) when prompts were created, `:no-library-content` when the Metabot has no models or metrics to summarize, or `:ai-produced-no-prompts` when generation produced nothing. Returns a 402 if the instance has reached its managed-AI usage limit.

    POST /api/metabot/metabot/{id}/prompt-suggestions/regenerate

  • Destructive

    DELETE /api/metabot/metabot/{id}/prompt-suggestions/{prompt-id}

    Delete the prompt suggestion with ID `prompt-id` for the metabot instance with `id`.

    DELETE /api/metabot/metabot/{id}/prompt-suggestions/{prompt-id}

  • Destructive

    PUT /api/metabot/slack/settings

    Update Metabot Slack settings atomically. All credential fields must be provided together. Setting values requires the metabot-v3 feature, but clearing values is always allowed. slack-connect-enabled is automatically set to true when credentials are provided, nil when cleared.

    PUT /api/metabot/slack/settings

  • Destructive

    POST /api/metric/{id}/dimension/remove

    Remove dimensions (by UUID) from a metric. Returns the updated list of added dimensions.

    POST /api/metric/{id}/dimension/remove

  • Destructive

    POST /api/metric/{id}/dimension/set-default

    Mark exactly one dimension as the metric's default, clearing any previous default. A null `dimension_id` clears the default without setting a new one, so the metric renders as a scalar. Returns the updated list of added dimensions.

    POST /api/metric/{id}/dimension/set-default

  • Destructive

    DELETE /api/model-index/{id}

    Delete ModelIndex.

    DELETE /api/model-index/{id}

  • Destructive

    PUT /api/mt/gtap/{id}

    Update a GTAP entry. The only things you're allowed to update for a GTAP are the Card being used (`card_id`) or the parameter mappings; changing `table_id` or `group_id` would effectively be deleting this entry and creating a new one. If that's what you want to do, do so explicitly with appropriate calls to the `DELETE` and `POST` endpoints.

    PUT /api/mt/gtap/{id}

  • Destructive

    DELETE /api/mt/gtap/{id}

    Delete a GTAP entry.

    DELETE /api/mt/gtap/{id}

  • Destructive

    POST /api/notification/admin/bulk

    Bulk-archive or -change-creator a set of notifications. The per-notification `:active` flip goes through `:model/Notification`'s `before-update` hook, which creates / tears down the Quartz triggers. Recipient emails and `:event/notification-update` audit events are published via the shared [[notification-api/publish-notification-update!]] helper so this endpoint's side-effect contract can't drift from `PUT /api/notification/:id`.

    POST /api/notification/admin/bulk

  • Destructive

    POST /api/notification/unsubscribe

    Allow non-users to unsubscribe from notifications, with the hash given through email.

    POST /api/notification/unsubscribe

  • Destructive

    POST /api/notification/{id}/unsubscribe

    Unsubscribe current user from a notification.

    POST /api/notification/{id}/unsubscribe

  • Destructive

    PUT /api/osi/ai-context/{entity-type}/{entity-local-id}

    Create or replace (upsert) the ai_context for an entity, addressed by its logical key. One row per entity: an existing row is updated in place rather than duplicated, and the compound primary key plus the upsert keep two concurrent writers from racing in a duplicate row.

    PUT /api/osi/ai-context/{entity-type}/{entity-local-id}

  • Destructive

    DELETE /api/osi/ai-context/{entity-type}/{entity-local-id}

    Delete the ai_context entry for an entity by its logical key.

    DELETE /api/osi/ai-context/{entity-type}/{entity-local-id}

  • Destructive

    DELETE /api/permissions/group/{group-id}

    Delete a specific `PermissionsGroup`.

    DELETE /api/permissions/group/{group-id}

  • Destructive

    PUT /api/permissions/membership/{group-id}/clear

    Remove all members from a `PermissionsGroup`. Returns a 400 (Bad Request) if the group ID is for the admin group.

    PUT /api/permissions/membership/{group-id}/clear

  • Destructive

    DELETE /api/permissions/membership/{id}

    Remove a User from a PermissionsGroup (delete their membership).

    DELETE /api/permissions/membership/{id}

  • Destructive

    POST /api/persist/card/{card-id}/unpersist

    Unpersist this model. Deletes the persisted table backing the model and all queries after this will use the card's query rather than the saved version of the query.

    POST /api/persist/card/{card-id}/unpersist

  • Destructive

    POST /api/persist/database/{id}/unpersist

    Attempt to disable model persistence for a database. If already not enabled, just returns a generic 204.

    POST /api/persist/database/{id}/unpersist

  • Destructive

    POST /api/persist/disable

    Disable global setting to allow databases to persist models. This will remove all tasks to refresh tables, remove that option from databases which might have it enabled, and delete all cached tables.

    POST /api/persist/disable

  • Destructive

    POST /api/premium-features/token/refresh

    Clear all token caches and re-check the premium features token against the MetaStore. Returns the fresh token status. Useful for the frontend after a purchase so that new features are recognized without waiting for the cache TTL.

    POST /api/premium-features/token/refresh

  • Destructive

    POST /api/pulse/unsubscribe

    Allow non-users to unsubscribe from pulses/subscriptions, with the hash given through email.

    POST /api/pulse/unsubscribe

  • Destructive

    DELETE /api/pulse/{id}/subscription

    For users to unsubscribe themselves from a pulse subscription.

    DELETE /api/pulse/{id}/subscription

  • Destructive

    POST /api/revision/revert

    Revert an object to a prior revision.

    POST /api/revision/revert

  • Destructive

    DELETE /api/segment/{id}

    Archive a Segment. (DEPRECATED -- Just pass updated value of `:archived` to the `PUT` endpoint instead.)

    DELETE /api/segment/{id}

  • Destructive

    DELETE /api/session

    Logout.

    DELETE /api/session

  • Destructive

    POST /api/session/forgot_password

    Send a reset email when user has forgotten their password.

    POST /api/session/forgot_password

  • Destructive

    POST /api/session/reset_password

    Reset password with a reset token.

    POST /api/session/reset_password

  • Destructive

    PUT /api/setting/{key}

    Create/update a `Setting`. If called by a non-admin, only user-local settings can be updated. This endpoint can also be used to delete Settings by passing `nil` for `:value`.

    PUT /api/setting/{key}

  • Destructive

    PUT /api/slack/settings

    Update Slack related settings. You must be a superuser to do this. Also updates the slack-cache. There are 3 cases where we alter the slack channel/user cache: 1. falsy token -> clear 2. invalid token -> clear 3. truthy, valid token -> refresh

    PUT /api/slack/settings

  • Destructive

    POST /api/table/{id}/discard_values

    Discard the FieldValues belonging to the Fields in this Table. Only applies to fields that have FieldValues. If this Table's Database is set up to automatically sync FieldValues, they will be recreated during the next cycle.

    POST /api/table/{id}/discard_values

  • Destructive

    POST /api/table/{id}/replace-csv

    Replaces the contents of the table identified by `:id` with the rows of an uploaded CSV file. The table must have been created by uploading a CSV file. The file may be at most 50 MB; larger uploads are rejected with a 413 response.

    POST /api/table/{id}/replace-csv

  • Destructive

    DELETE /api/timeline-event/{id}

    Delete a [[TimelineEvent]].

    DELETE /api/timeline-event/{id}

  • Destructive

    PUT /api/timeline/{id}

    Update the [[Timeline]] with `id`. Returns the timeline without events. Archiving a timeline will archive all of the events in that timeline.

    PUT /api/timeline/{id}

  • Destructive

    DELETE /api/timeline/{id}

    Delete a [[Timeline]]. Will cascade delete its events as well.

    DELETE /api/timeline/{id}

  • Destructive

    POST /api/transform-dag-run/{run-id}/cancel

    Cancel an in-progress manual DAG run and request cancellation of its still-running transforms.

    POST /api/transform-dag-run/{run-id}/cancel

  • Destructive

    DELETE /api/transform-job/{job-id}

    Delete a transform job.

    DELETE /api/transform-job/{job-id}

  • Destructive

    POST /api/transform-job/{job-id}/runs/{run-id}/cancel

    Cancel an in-progress job run and request cancellation of its still-running transforms.

    POST /api/transform-job/{job-id}/runs/{run-id}/cancel

  • Destructive

    DELETE /api/transform-tag/{tag-id}

    Delete a transform tag. Removes it from all transforms and jobs.

    DELETE /api/transform-tag/{tag-id}

  • Destructive

    DELETE /api/transform/{id}

    Delete a transform.

    DELETE /api/transform/{id}

  • Destructive

    POST /api/transform/{id}/cancel

    Cancel the current run for a given transform.

    POST /api/transform/{id}/cancel

  • Destructive

    POST /api/transform/{id}/reset-checkpoint

    Reset the stored checkpoint for an incremental transform.

    POST /api/transform/{id}/reset-checkpoint

  • Destructive

    DELETE /api/transform/{id}/table

    Delete a transform's output table.

    DELETE /api/transform/{id}/table

  • Destructive

    DELETE /api/user-key-value/namespace/{namespace}/key/{key}

    Deletes a KV-pair for the user

    DELETE /api/user-key-value/namespace/{namespace}/key/{key}

  • Destructive

    PUT /api/user/{id}

    Update an existing, active `User`. Self or superusers can update user info and groups. Group Managers can only add/remove users from groups they are manager of. Only personal users can be updated this way; API-key users 404 (manage them via `/api/api-key` instead).

    PUT /api/user/{id}

  • Destructive

    DELETE /api/user/{id}

    Disable a `User`. This does not remove the `User` from the DB, but instead disables their account.

    DELETE /api/user/{id}

  • Destructive

    POST /api/user/{id}/password-reset-url

    Generate a password reset URL for a user. Admins can share this URL directly with the user. The link expires in 48 hours.

    POST /api/user/{id}/password-reset-url

See How connections work for access and permissions, or browse all integrations.

Reference date: October 29, 2026. Metabase API docs.