Skip to content
Changelog

Report a bug

Hyperbrowser integration: operations for AI agents

Hyperbrowser connects to Type.com as an integration. This page lists the 111 operations documented for Claude and Codex in Type.com: 55 read, 33 write, 23 destructive.

About Hyperbrowser

Hyperbrowser is a next-generation platform empowering AI agents and enabling effortless, scalable browser automation.

Categories: Analytics & Data

How to connect Hyperbrowser to Type.com

Access
Organization connection: one shared account for the workspace.

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

Documented operations

111 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.

111 operations listed.

Session

  • Read

    Get session by ID

    GET /api/session/{id}

  • Read

    Get session recording URL

    GET /api/session/{id}/recording-url

  • Read

    Get session video recording URL

    GET /api/session/{id}/video-recording-url

  • Read

    Get session downloads URL

    GET /api/session/{id}/downloads-url

  • Read

    List WebMCP tools

    Discover tools across open tabs and frames. Inspect source information as tool names can repeat. Tool references are bound to their original document. Requires an active session created with enableWebMcp: true and an API key with browser.sessions.update permission.

  • Read

    Get a WebMCP invocation result

    Read or long-poll a handle without re-invoking the tool. A running or awaitingsubmission response is still pending. A waitSeconds timeout does not cancel the invocation.

  • Write

    Create new session

    POST /api/session

  • Write

    Run manual CAPTCHA evaluation

    Trigger a bounded CAPTCHA evaluation on an active browser session. Supported CAPTCHA targets are turnstile, cloudflare-challenge, aliexpress, recaptcha, and amazon.

  • Write

    Invoke a WebMCP tool

    Invoke once and wait for the result. A form may return awaitingsubmission with an invocationId. Do not blindly retry if the response is lost; the tool may already have acted.

  • Write

    Start a WebMCP invocation

    Start once and return a handle without waiting for completion. Acceptance does not guarantee success; inspect the handle later. A lost response may mean losing the handle to a tool that already acted. Do not blindly retry.

  • Write

    Fill a credential

    Types a secret, a vault's entries, or fields read from 1Password into inputs on a page in the session. The values are decrypted on the session's host; neither the request nor the response contains them.

  • Destructive

    Stop a session

    PUT /api/session/{id}/stop

  • Destructive

    Update a running session

    Update supported settings on an active browser session. Supported update types are profile, proxy, screen, and solveCaptchas.

  • Destructive

    Cancel a WebMCP invocation

    Request best-effort cancellation and return the current handle. Completion can win the race. This does not roll back page side effects. Continue result reads if still pending.

Sessions

  • Read

    Get list of sessions

    GET /api/sessions

Profile

  • Read

    Get profile by ID

    GET /api/profile/{id}

  • Write

    Creates a new profile

    POST /api/profile

  • Destructive

    Delete profile by ID

    DELETE /api/profile/{id}

Profiles

  • Read

    Get list of profiles

    GET /api/profiles

Scrape

  • Read

    Get scrape job status and result

    GET /api/scrape/{id}

  • Read

    Get scrape job status

    GET /api/scrape/{id}/status

  • Read

    Get batch scrape job status and results

    GET /api/scrape/batch/{id}

  • Read

    Get batch scrape job status

    GET /api/scrape/batch/{id}/status

  • Write

    Create new scrape job

    POST /api/scrape

  • Write

    Start a batch scrape job

    POST /api/scrape/batch

Crawl

  • Read

    Get crawl job status and results

    GET /api/crawl/{id}

  • Read

    Get crawl job status

    GET /api/crawl/{id}/status

  • Write

    Start a crawl job

    POST /api/crawl

Extract

  • Read

    Get extract job status and results

    GET /api/extract/{id}

  • Read

    Get extract job status

    GET /api/extract/{id}/status

  • Write

    Start an extract job

    POST /api/extract

Task

  • Read

    Get HyperAgent task status and results

    GET /api/task/hyper-agent/{id}

  • Read

    Get HyperAgent task status

    GET /api/task/hyper-agent/{id}/status

  • Read

    Get browser use task status and results

    GET /api/task/browser-use/{id}

  • Read

    Get browser use task status

    GET /api/task/browser-use/{id}/status

  • Read

    Get claude computer use task status and results

    GET /api/task/claude-computer-use/{id}

  • Read

    Get claude computer use task status

    GET /api/task/claude-computer-use/{id}/status

  • Read

    Get gemini computer use task status and results

    GET /api/task/gemini-computer-use/{id}

  • Read

    Get gemini computer use task status

    GET /api/task/gemini-computer-use/{id}/status

  • Read

    Get meta computer use task status and results

    GET /api/task/meta-computer-use/{id}

  • Read

    Get meta computer use task status

    GET /api/task/meta-computer-use/{id}/status

  • Read

    Get grok computer use task status and results

    GET /api/task/grok-computer-use/{id}

  • Read

    Get grok computer use task status

    GET /api/task/grok-computer-use/{id}/status

  • Read

    Get CUA task status and results

    GET /api/task/cua/{id}

  • Read

    Get CUA task status

    GET /api/task/cua/{id}/status

  • Read

    Get jev computer use task status and results

    GET /api/task/jev/{id}

  • Read

    Get jev computer use task status

    GET /api/task/jev/{id}/status

  • Write

    Start a HyperAgent task

    POST /api/task/hyper-agent

  • Write

    Start a browser use task

    POST /api/task/browser-use

  • Write

    Start a claude computer use task

    POST /api/task/claude-computer-use

  • Write

    Start a gemini computer use task

    POST /api/task/gemini-computer-use

  • Write

    Start a meta computer use task

    POST /api/task/meta-computer-use

  • Write

    Start a grok computer use task

    POST /api/task/grok-computer-use

  • Write

    Start a CUA task

    POST /api/task/cua

  • Write

    Start a jev computer use task

    POST /api/task/jev

  • Destructive

    Stop a HyperAgent task

    PUT /api/task/hyper-agent/{id}/stop

  • Destructive

    Stop a browser use task

    PUT /api/task/browser-use/{id}/stop

  • Destructive

    Stop a claude computer use task

    PUT /api/task/claude-computer-use/{id}/stop

  • Destructive

    Stop a gemini computer use task

    PUT /api/task/gemini-computer-use/{id}/stop

  • Destructive

    Stop a meta computer use task

    PUT /api/task/meta-computer-use/{id}/stop

  • Destructive

    Stop a grok computer use task

    PUT /api/task/grok-computer-use/{id}/stop

  • Destructive

    Stop a CUA task

    PUT /api/task/cua/{id}/stop

  • Destructive

    Stop a jev computer use task

    PUT /api/task/jev/{id}/stop

Extensions

  • Read

    List all extensions

    GET /api/extensions/list

  • Write

    Add a new extension

    POST /api/extensions/add

Web

  • Read

    Fetch a web page

    Fetches a web page and returns the content in various formats (HTML, Markdown, JSON, screenshot, etc.)

  • Read

    Search the web

    Performs a web search and returns search results with titles, URLs, and descriptions

  • Read

    Get web crawl job results

    Retrieves the status and results of a web crawl job. Results are paginated.

  • Read

    Get web crawl job status

    Retrieves just the status of a web crawl job without the full results.

  • Write

    Start a web crawl job

    Starts an asynchronous crawl job that follows links from a starting URL and returns content from each page in the specified formats.

X402

  • Read

    Fetch a web page with X402 payment

    X402 payment endpoint. First request returns 402 with payment requirements. Retry with PAYMENT-SIGNATURE header containing cryptographic proof to get the actual data. See https://x402.gitbook.io/x402 for protocol details.

  • Read

    Search the web with X402 payment

    X402 payment endpoint. First request returns 402 with payment requirements. Retry with PAYMENT-SIGNATURE header containing cryptographic proof to get search results. See https://x402.gitbook.io/x402 for protocol details.

  • Read

    Get an X402 session

    Get the purchased session's status and original expiry without another payment. Authenticate with managementToken, not the browser token. Connection credentials are not reissued.

  • Write

    Create a session with X402 payment

    Purchase one browser session at 0.10 USDC per hour, paid upfront for timeoutMinutes (5–60, default 30). No Hyperbrowser account or API key is required. A valid unpaid request returns 402 with payment requirements. Retry with a signed PAYMENT-SIGNATURE authorization.

  • Destructive

    Stop an X402 session

    Stop the purchased browser without another payment. Authenticate with managementToken, not the browser token. An already closed or errored session also succeeds.

Sandboxes

  • Read

    Get sandbox by ID

    GET /api/sandbox/{id}

  • Read

    Get list of sandboxes

    GET /api/sandboxes

  • Write

    Create new sandbox

    Create a CPU, GPU, or macOS sandbox. For GPU creation, set gpu and use a GPU-ready team image with region us. GPU access must be enabled for your team; contact info@hyperbrowser.ai. GPU requests return HTTP 202 with status provisioning and null runtime and token.

  • Write

    Expose a sandbox port

    POST /api/sandbox/{id}/expose

  • Write

    Create a sandbox memory snapshot

    POST /api/sandbox/{id}/snapshot

  • Write

    Start a sandbox process with secrets

    Starts a process in a running sandbox with secrets as environment variables for that process only.

  • Destructive

    Stop a sandbox

    PUT /api/sandbox/{id}/stop

  • Destructive

    Remove an exposed sandbox port

    POST /api/sandbox/{id}/unexpose

  • Destructive

    Update sandbox network policy

    PUT /api/sandbox/{id}/network

Volumes

  • Read

    Get list of volumes

    GET /api/volume

  • Read

    Get volume by ID

    GET /api/volume/{id}

  • Write

    Create a volume

    POST /api/volume

  • Destructive

    Delete a volume

    DELETE /api/volume/{key}

Images

  • Read

    Get list of sandbox images

    GET /api/images

  • Read

    Get list of image builds

    GET /api/images/builds

  • Read

    Get image build by ID

    GET /api/images/builds/{buildId}

  • Write

    Create an image build

    POST /api/images/builds

  • Write

    Complete an image build upload

    POST /api/images/builds/{buildId}/complete

  • Destructive

    Delete sandbox image

    DELETE /api/images/{key}

  • Destructive

    Cancel an image build

    POST /api/images/builds/{buildId}/cancel

Snapshots

  • Read

    Get list of sandbox snapshots

    GET /api/snapshots

  • Read

    Get sandbox snapshot

    GET /api/snapshots/{key}

  • Destructive

    Delete sandbox snapshot

    DELETE /api/snapshots/{key}

Secrets

  • Read

    Get the secrets public key

    Returns the RSA public key that you encrypt secret values to before uploading them with POST /api/secrets/{id}/versions. Doesn't require authentication.

  • Read

    List secrets

    Lists your team's secrets, sorted by name. Values are never returned. Secrets are enabled for teams on request; please contact support to enable them (otherwise 403 SECRETSFEATUREDISABLED). Restricted API keys need the secrets.read scope.

  • Read

    Get a secret

    Returns a secret and its stored versions. Values are never returned. Secrets are enabled for teams on request; please contact support to enable them (otherwise 403 SECRETSFEATUREDISABLED). Restricted API keys need the secrets.read scope.

  • Write

    Create a secret

    Creates a secret without a value. Then upload its value with POST /api/secrets/{id}/versions, encrypted with the encryptionContext this returns. A team can have at most 100 secrets.

  • Write

    Update a secret

    Updates a secret's description or metadata, such as its allowed origins, without uploading a new value. Metadata is merged by top-level member: each member you send replaces the stored one, and null removes it.

  • Write

    Add a secret version

    Uploads a new value as the secret's next version. Encrypt the value on your side as a compact JWE (RSA-OAEP-256 and A256GCM) to the key from GET /api/secrets/public-key, with the protected header {alg, enc, kid, ctx}, where ctx is the secret's encryptionContext.

  • Destructive

    Delete a secret

    Permanently deletes a secret and all its versions. A secret that a vault uses as an entry can't be deleted. Secrets are enabled for teams on request; please contact support to enable them (otherwise 403 SECRETSFEATUREDISABLED). Restricted API keys need the secrets.manage scope.

  • Destructive

    Delete a secret version

    Permanently deletes an older version of a secret. You can't delete the latest version: add a new version or delete the secret instead. Secrets are enabled for teams on request; please contact support to enable them (otherwise 403 SECRETSFEATUREDISABLED).

Vaults

  • Read

    List vaults

    Lists your team's vaults, sorted by name, with their entries. Secrets are enabled for teams on request; please contact support to enable them (otherwise 403 SECRETSFEATUREDISABLED). Restricted API keys need the secrets.read scope.

  • Read

    Get a vault

    Returns a vault and its entries. Secrets are enabled for teams on request; please contact support to enable them (otherwise 403 SECRETSFEATUREDISABLED). Restricted API keys need the secrets.read scope.

  • Write

    Create a vault

    Creates a vault: a set of environment variables that point at your secrets. Pass the vault by name as vault to POST /api/sandbox and POST /api/sandbox/{id}/processes, or as {"vault": name} in a browser session's credentials. A team can have at most 100 vaults.

  • Write

    Update a vault

    Changes a vault's description or entries. Entries you pass replace those variables and null removes one; others are kept. Processes that already started keep the values they received.

  • Destructive

    Delete a vault

    Permanently deletes a vault. The secrets it points at are kept. Secrets are enabled for teams on request; please contact support to enable them (otherwise 403 SECRETSFEATUREDISABLED). Restricted API keys need the secrets.manage scope.

See How connections work for access and permissions, or browse all integrations.