Steel Scrape API

openapi: 3.0.1
info:
  title: Steel API
  description: >-
    Steel is the open-source browser API for AI agents and apps. The Steel
    Cloud REST API launches and manages cloud browser sessions, runs stateless
    quick actions (scrape, screenshot, pdf, search), and exposes a live session
    viewer. Long-running automation connects to the per-session Chrome
    DevTools Protocol (CDP) WebSocket returned as `websocketUrl`, which is
    driven with Playwright, Puppeteer, or Selenium. The same surface is
    available self-hosted (Apache-2.0) from the steel-browser server, where the
    default base path is http://localhost:3000/v1.
  termsOfService: https://steel.dev/terms-of-service
  contact:
    name: Steel Support
    url: https://docs.steel.dev
  license:
    name: Apache 2.0
    url: https://github.com/steel-dev/steel-browser/blob/main/LICENSE
  version: '1.0'
servers:
  - url: https://api.steel.dev/v1
    description: Steel Cloud
  - url: http://localhost:3000/v1
    description: Self-hosted steel-browser
security:
  - SteelApiKey: []
tags:
  - name: Sessions
    description: Launch, inspect, and release cloud browser sessions.
  - name: Session Actions
    description: Scrape, screenshot, and PDF the page in a running session.
  - name: Quick Actions
    description: Stateless one-off scrape, screenshot, PDF, and search.
  - name: Files
    description: Manage files inside a session's browser context.
paths:
  /sessions:
    post:
      operationId: createSession
      tags:
        - Sessions
      summary: Create a browser session
      description: >-
        Launch a new cloud browser session with optional proxy, fingerprint,
        dimensions, timezone, ad-blocking, and bandwidth options. Returns the
        session details including the `websocketUrl` (CDP) used to connect
        Playwright/Puppeteer over CDP, plus the live session viewer URLs.
      requestBody:
        required: false
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CreateSession'
      responses:
        '200':
          description: Session created
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/SessionDetails'
    get:
      operationId: listSessions
      tags:
        - Sessions
      summary: List sessions
      description: Returns all sessions for the authenticated account.
      responses:
        '200':
          description: A list of sessions
          content:
            application/json:
              schema:
                type: object
                properties:
                  sessions:
                    type: array
                    items:
                      $ref: '#/components/schemas/SessionDetails'
  /sessions/{sessionId}:
    get:
      operationId: getSession
      tags:
        - Sessions
      summary: Get session details
      parameters:
        - $ref: '#/components/parameters/SessionId'
      responses:
        '200':
          description: Session details
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/SessionDetails'
  /sessions/{sessionId}/context:
    get:
      operationId: getSessionContext
      tags:
        - Sessions
      summary: Get session context
      description: >-
        Returns the browser context (cookies, localStorage, sessionStorage,
        IndexedDB) captured for the session, suitable for persisting and
        replaying authenticated state.
      parameters:
        - $ref: '#/components/parameters/SessionId'
      responses:
        '200':
          description: Session context
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/SessionContext'
  /sessions/{sessionId}/live-details:
    get:
      operationId: getSessionLiveDetails
      tags:
        - Sessions
      summary: Get session live details
      description: >-
        Returns the live session viewer URLs, the CDP `websocketUrl`, the list
        of open pages/tabs, and the current browser state.
      parameters:
        - $ref: '#/components/parameters/SessionId'
      responses:
        '200':
          description: Live session details
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/SessionLiveDetails'
  /sessions/{sessionId}/release:
    post:
      operationId: releaseSession
      tags:
        - Sessions
      summary: Release a session
      description: Releases (ends) a single browser session and frees its resources.
      parameters:
        - $ref: '#/components/parameters/SessionId'
      responses:
        '200':
          description: Session released
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ReleaseSession'
  /sessions/release:
    post:
      operationId: releaseSessions
      tags:
        - Sessions
      summary: Release all sessions
      description: Releases all active browser sessions for the account.
      responses:
        '200':
          description: Sessions released
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ReleaseSession'
  /sessions/scrape:
    post:
      operationId: scrapeSession
      tags:
        - Session Actions
      summary: Scrape the current session page
      description: Extracts content from the page currently loaded in the session.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/SessionScrapeRequest'
      responses:
        '200':
          description: Scrape result
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ScrapeResponse'
  /sessions/screenshot:
    post:
      operationId: screenshotSession
      tags:
        - Session Actions
      summary: Screenshot the current session page
      requestBody:
        required: false
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/SessionScreenshotRequest'
      responses:
        '200':
          description: PNG screenshot
          content:
            image/png:
              schema:
                type: string
                format: binary
  /sessions/pdf:
    post:
      operationId: pdfSession
      tags:
        - Session Actions
      summary: Generate a PDF of the current session page
      requestBody:
        required: false
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/SessionPDFRequest'
      responses:
        '200':
          description: PDF document
          content:
            application/pdf:
              schema:
                type: string
                format: binary
  /scrape:
    post:
      operationId: scrape
      tags:
        - Quick Actions
      summary: Scrape a URL
      description: >-
        Stateless one-off scrape. Loads the URL in a fresh browser and returns
        the requested formats (html, readability, cleaned_html, markdown) plus
        page metadata and links, with optional screenshot and PDF.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ScrapeRequest'
      responses:
        '200':
          description: Scrape result
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ScrapeResponse'
  /screenshot:
    post:
      operationId: screenshot
      tags:
        - Quick Actions
      summary: Take a screenshot of a URL
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ScreenshotRequest'
      responses:
        '200':
          description: PNG screenshot
          content:
            image/png:
              schema:
                type: string
                format: binary
  /pdf:
    post:
      operationId: pdf
      tags:
        - Quick Actions
      summary: Get the PDF of a URL
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/PDFRequest'
      responses:
        '200':
          description: PDF document
          content:
            application/pdf:
              schema:
                type: string
                format: binary
  /search:
    post:
      operationId: search
      tags:
        - Quick Actions
      summary: Search the web
      description: Uses a search engine to return URLs for a query.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/SearchRequest'
      responses:
        '200':
          description: Search results
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/SearchResponse'
  /sessions/{sessionId}/files:
    get:
      operationId: listSessionFiles
      tags:
        - Files
      summary: List files in a session
      parameters:
        - $ref: '#/components/parameters/SessionId'
      responses:
        '200':
          description: List of files
          content:
            application/json:
              schema:
                type: object
                properties:
                  data:
                    type: array
                    items:
                      $ref: '#/components/schemas/SessionFile'
    post:
      operationId: uploadSessionFile
      tags:
        - Files
      summary: Upload a file to a session
      parameters:
        - $ref: '#/components/parameters/SessionId'
      requestBody:
        required: true
        content:
          multipart/form-data:
            schema:
              type: object
              properties:
                file:
                  type: string
                  format: binary
                path:
                  type: string
      responses:
        '200':
          description: Uploaded file
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/SessionFile'
    delete:
      operationId: deleteSessionFiles
      tags:
        - Files
      summary: Delete all files in a session
      parameters:
        - $ref: '#/components/parameters/SessionId'
      responses:
        '200':
          description: Files deleted
          content:
            application/json:
              schema:
                type: object
                properties:
                  success:
                    type: boolean
components:
  securitySchemes:
    SteelApiKey:
      type: apiKey
      in: header
      name: Steel-Api-Key
      description: >-
        API key issued from app.steel.dev. Pass it in the `Steel-Api-Key`
        request header. Self-hosted instances may run without auth.
  parameters:
    SessionId:
      name: sessionId
      in: path
      required: true
      description: Unique identifier of the session.
      schema:
        type: string
        format: uuid
  schemas:
    CreateSession:
      type: object
      description: Options for launching a browser session.
      properties:
        sessionId:
          type: string
          format: uuid
          description: Optional client-supplied session identifier.
        proxyUrl:
          type: string
          description: Proxy URL to use for the session.
        userAgent:
          type: string
          description: User agent string to use for the session.
        sessionContext:
          $ref: '#/components/schemas/SessionContext'
        blockAds:
          type: boolean
          description: Block ads in the session.
        solveCaptcha:
          type: boolean
          description: Enable automatic CAPTCHA solving.
        optimizeBandwidth:
          description: >-
            Enable bandwidth optimizations. `true` enables all flags; an object
            allows granular control over blocked resource types and hosts.
          oneOf:
            - type: boolean
            - type: object
              properties:
                blockImages:
                  type: boolean
                blockMedia:
                  type: boolean
                blockStylesheets:
                  type: boolean
                blockHosts:
                  type: array
                  items:
                    type: string
                blockUrlPatterns:
                  type: array
                  items:
                    type: string
        skipFingerprintInjection:
          type: boolean
          description: Skip fingerprint injection for this session.
        deviceConfig:
          type: object
          properties:
            device:
              type: string
              enum: [desktop, mobile]
              default: desktop
        fullscreen:
          type: boolean
          description: Launch the browser in fullscreen mode with no Chrome UI.
        extensions:
          type: array
          items:
            type: string
          description: Browser extensions to load.
        persist:
          type: boolean
          description: Persist the session for later resumption.
        timezone:
          type: string
          description: Timezone to use for the session.
        dimensions:
          $ref: '#/components/schemas/Dimensions'
        credentials:
          type: object
          description: Configuration for autofilled session credentials.
          properties:
            autoSubmit:
              type: boolean
            blurFields:
              type: boolean
            exactOrigin:
              type: boolean
        isSelenium:
          type: boolean
          description: Indicates if the session is driven over Selenium.
    SessionDetails:
      type: object
      required:
        - id
        - createdAt
        - status
        - websocketUrl
      properties:
        id:
          type: string
          format: uuid
        createdAt:
          type: string
          format: date-time
        status:
          type: string
          enum: [idle, live, released, failed]
        duration:
          type: integer
          description: Duration of the session in milliseconds.
        eventCount:
          type: integer
        timeout:
          type: integer
          description: Session timeout in milliseconds.
        creditsUsed:
          type: integer
          description: Credits consumed by the session.
        websocketUrl:
          type: string
          description: >-
            CDP WebSocket URL (wss://) for the session. Connect with
            Playwright/Puppeteer via connect_over_cdp / connectOverCDP.
        debugUrl:
          type: string
          description: URL for viewing the live browser instance.
        debuggerUrl:
          type: string
          description: URL for debugging the session.
        sessionViewerUrl:
          type: string
          description: Live session viewer URL.
        dimensions:
          $ref: '#/components/schemas/Dimensions'
        userAgent:
          type: string
        proxy:
          type: string
        proxyTxBytes:
          type: integer
        proxyRxBytes:
          type: integer
        solveCaptcha:
          type: boolean
        isSelenium:
          type: boolean
    ReleaseSession:
      allOf:
        - $ref: '#/components/schemas/SessionDetails'
        - type: object
          properties:
            success:
              type: boolean
    SessionLiveDetails:
      type: object
      properties:
        sessionViewerUrl:
          type: string
        sessionViewerFullscreenUrl:
          type: string
        websocketUrl:
          type: string
          description: CDP WebSocket URL for the session.
        pages:
          type: array
          items:
            type: object
            properties:
              id:
                type: string
              url:
                type: string
              title:
                type: string
              favicon:
                type: string
                nullable: true
        browserState:
          type: object
          properties:
            status:
              type: string
              enum: [idle, live, released, failed]
            userAgent:
              type: string
            browserVersion:
              type: string
            initialDimensions:
              $ref: '#/components/schemas/Dimensions'
            pageCount:
              type: integer
    SessionContext:
      type: object
      description: Captured browser state for the session.
      properties:
        cookies:
          type: array
          items:
            type: object
            additionalProperties: true
        localStorage:
          type: object
          additionalProperties: true
        sessionStorage:
          type: object
          additionalProperties: true
        indexedDB:
          type: object
          additionalProperties: true
    Dimensions:
      type: object
      properties:
        width:
          type: integer
        height:
          type: integer
    SessionScrapeRequest:
      type: object
      description: Scrape request for the current session page (no url).
      properties:
        format:
          type: array
          items:
            $ref: '#/components/schemas/ScrapeFormat'
        screenshot:
          type: boolean
        pdf:
          type: boolean
        removeBase64Images:
          type: boolean
        proxyUrl:
          type: string
          nullable: true
        delay:
          type: number
    SessionScreenshotRequest:
      type: object
      properties:
        fullPage:
          type: boolean
        proxyUrl:
          type: string
          nullable: true
        delay:
          type: number
    SessionPDFRequest:
      type: object
      properties:
        proxyUrl:
          type: string
          nullable: true
        delay:
          type: number
    ScrapeFormat:
      type: string
      enum:
        - html
        - readability
        - cleaned_html
        - markdown
    ScrapeRequest:
      type: object
      required:
        - url
      properties:
        url:
          type: string
        format:
          type: array
          items:
            $ref: '#/components/schemas/ScrapeFormat'
        screenshot:
          type: boolean
        pdf:
          type: boolean
        removeBase64Images:
          type: boolean
        proxyUrl:
          type: string
          nullable: true
        delay:
          type: number
    ScrapeResponse:
      type: object
      properties:
        content:
          type: object
          description: Map of requested format to extracted content.
          additionalProperties: true
        metadata:
          type: object
          properties:
            title:
              type: string
            language:
              type: string
            urlSource:
              type: string
            timestamp:
              type: string
              format: date-time
            description:
              type: string
            statusCode:
              type: integer
            wordCount:
              type: number
            canonical:
              type: string
            favicon:
              type: string
        links:
          type: array
          items:
            type: object
            properties:
              url:
                type: string
              text:
                type: string
        screenshot:
          type: string
          description: Base64 screenshot, present when screenshot=true.
        pdf:
          type: string
          description: Base64 PDF, present when pdf=true.
    ScreenshotRequest:
      type: object
      required:
        - url
      properties:
        url:
          type: string
        fullPage:
          type: boolean
        proxyUrl:
          type: string
          nullable: true
        delay:
          type: number
    PDFRequest:
      type: object
      required:
        - url
      properties:
        url:
          type: string
        proxyUrl:
          type: string
          nullable: true
        delay:
          type: number
    SearchRequest:
      type: object
      required:
        - query
      properties:
        query:
          type: string
        proxyUrl:
          type: string
          nullable: true
    SearchResponse:
      type: object
      properties:
        results:
          type: array
          items:
            type: object
            properties:
              title:
                type: string
              url:
                type: string
              description:
                type: string
                nullable: true
    SessionFile:
      type: object
      properties:
        path:
          type: string
        size:
          type: integer
        lastModified:
          type: string
          format: date-time