openapi: 3.1.0
info:
  title: Parallel Chat API (Beta) Chat API (Beta) Extract API
  description: Parallel API
  contact:
    name: Parallel Support
    url: https://parallel.ai
    email: support@parallel.ai
  version: 0.1.2
servers:
- url: https://api.parallel.ai
  description: Parallel API
security:
- ApiKeyAuth: []
tags:
- name: Extract
  description: 'Extract returns excerpts or full content from one or more URLs. Inputs are a list of URLs and an optional search objective and keyword queries. The returned excerpts or full content is formatted as markdown and suitable for LLM consumption.

    - Result: excerpts or full content from the URL formatted as markdown'
paths:
  /v1/extract:
    post:
      tags:
      - Extract
      summary: Extract
      description: 'Extracts relevant content from specific web URLs.


        The legacy Extract API reference (`/v1beta/extract` endpoint) is available

        [here](https://docs.parallel.ai/api-reference/legacy/extract-beta/extract), and

        migration guide is [here](https://docs.parallel.ai/extract/extract-migration-guide).'
      operationId: extract_v1_extract_post
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/V1ExtractRequest'
        required: true
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/V1ExtractResponse'
              example:
                extract_id: extract_8a911eb27c7a4afaa20d0d9dc98d07c0
                results:
                - url: https://www.example.com
                  title: Example Title
                  excerpts:
                  - Excerpted text ...
                  full_content: Full content ...
                errors:
                - url: https://www.example.com
                  error_type: fetch_error
                  http_status_code: 500
                  content: Error fetching content from https://www.example.com
                session_id: session_8a911eb27c7a4afaa20d0d9dc98d07c0
        '422':
          description: Request validation error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
              example:
                type: error
                error:
                  ref_id: extract_8a911eb27c7a4afaa20d0d9dc98d07c0
                  message: Request validation error
      x-code-samples:
      - lang: Python
        source: "from parallel import Parallel\n\nclient = Parallel()\n\nextract = client.extract(\n    urls=[\"https://www.example.com\"],\n    objective=\"Summarize the page\",\n)\nprint(extract.results)"
      - lang: TypeScript
        source: "import Parallel from \"parallel-web\";\n\nconst client = new Parallel();\n\nconst extract = await client.extract({\n    urls: [\"https://www.example.com\"],\n    objective: \"Summarize the page\",\n});\nconsole.log(extract.results);"
      - lang: cURL
        source: "curl --request POST \\\n    --url https://api.parallel.ai/v1/extract \\\n    --header 'Content-Type: application/json' \\\n    --header 'x-api-key: <api-key>' \\\n    --data '{\n    \"urls\": [\"https://www.example.com\"],\n    \"objective\": \"Summarize the page\"\n}'"
components:
  schemas:
    UsageItem:
      properties:
        name:
          type: string
          title: Name
          description: Name of the SKU.
          examples:
          - sku_search_additional_results
          - sku_extract_excerpts
        count:
          type: integer
          title: Count
          description: Count of the SKU.
          examples:
          - 1
      type: object
      required:
      - name
      - count
      title: UsageItem
      description: Usage item for a single operation.
    V1ExtractResult:
      properties:
        url:
          type: string
          title: Url
          description: URL associated with the search result.
        title:
          anyOf:
          - type: string
          - type: 'null'
          title: Title
          description: Title of the webpage, if available.
        publish_date:
          anyOf:
          - type: string
          - type: 'null'
          title: Publish Date
          description: Publish date of the webpage in YYYY-MM-DD format, if available.
        excerpts:
          items:
            type: string
          type: array
          title: Excerpts
          description: Relevant excerpted content from the URL, formatted as markdown.
        full_content:
          anyOf:
          - type: string
          - type: 'null'
          title: Full Content
          description: Full content from the URL formatted as markdown, if requested.
      type: object
      required:
      - url
      - excerpts
      title: V1ExtractResult
      description: Extract result for a single URL.
    FullContentSettings:
      properties:
        max_chars_per_result:
          anyOf:
          - type: integer
          - type: 'null'
          title: Max Chars Per Result
          description: Optional limit on the number of characters to include in the full content for each url. Full content always starts at the beginning of the page and is truncated at the limit if necessary.
      type: object
      title: FullContentSettings
      description: Optional settings for returning full content.
    ErrorResponse:
      properties:
        type:
          type: string
          const: error
          title: Type
          description: Always 'error'.
        error:
          $ref: '#/components/schemas/Error'
          description: Error.
      type: object
      required:
      - type
      - error
      title: ErrorResponse
      description: Response object used for non-200 status codes.
    ExtractError:
      properties:
        url:
          type: string
          title: Url
        error_type:
          type: string
          title: Error Type
          description: Error type.
        http_status_code:
          anyOf:
          - type: integer
          - type: 'null'
          title: Http Status Code
          description: HTTP status code, if available.
        content:
          anyOf:
          - type: string
          - type: 'null'
          title: Content
          description: Content returned for http client or server errors, if any.
      type: object
      required:
      - url
      - error_type
      - http_status_code
      - content
      title: ExtractError
      description: Extract error details.
    V1ExcerptSettings:
      properties:
        max_chars_per_result:
          anyOf:
          - type: integer
          - type: 'null'
          title: Max Chars Per Result
          description: Optional upper bound on the total number of characters to include per url. Excerpts may contain fewer characters than this limit to maximize relevance and token efficiency.
      additionalProperties: false
      type: object
      title: V1ExcerptSettings
      description: Optional settings for returning relevant excerpts.
    AdvancedExtractSettings:
      properties:
        fetch_policy:
          anyOf:
          - $ref: '#/components/schemas/FetchPolicy'
          - type: 'null'
          description: 'Fetch policy: determines when to return cached content from the index (faster) vs fetching live content (fresher). Default is to use a dynamic policy based on the search objective and url. Note: enabling live fetch significantly increases extract latency because it requires fetching content from source websites.'
        excerpt_settings:
          anyOf:
          - $ref: '#/components/schemas/V1ExcerptSettings'
          - type: 'null'
          description: Controls excerpt sizes. Provide excerpt settings for fine-grained control, or omit to use defaults.
        full_content:
          anyOf:
          - $ref: '#/components/schemas/FullContentSettings'
          - type: boolean
          title: Full Content
          description: Controls full content extraction. Set to true to enable with defaults, false to disable, or provide FullContentSettings for fine-grained control.
          default: false
      additionalProperties: false
      type: object
      title: AdvancedExtractSettings
      description: 'Advanced extract configuration.


        These settings may impact result quality and latency unless used carefully.

        See https://docs.parallel.ai/search/advanced-extract-settings for more info.'
    Warning:
      properties:
        type:
          type: string
          enum:
          - spec_validation_warning
          - input_validation_warning
          - warning
          title: Type
          description: Type of warning. Note that adding new warning types is considered a backward-compatible change.
          examples:
          - spec_validation_warning
          - input_validation_warning
        message:
          type: string
          title: Message
          description: Human-readable message.
        detail:
          anyOf:
          - additionalProperties: true
            type: object
          - type: 'null'
          title: Detail
          description: Optional detail supporting the warning.
      type: object
      required:
      - type
      - message
      title: Warning
      description: Human-readable message for a task.
    FetchPolicy:
      properties:
        max_age_seconds:
          anyOf:
          - type: integer
          - type: 'null'
          title: Max Age Seconds
          description: Maximum age of cached content in seconds to trigger a live fetch. Minimum value 600 seconds (10 minutes).
          examples:
          - 86400
        timeout_seconds:
          anyOf:
          - type: number
          - type: 'null'
          title: Timeout Seconds
          description: Timeout in seconds for fetching live content if unavailable in cache.
          examples:
          - 60
        disable_cache_fallback:
          type: boolean
          title: Disable Cache Fallback
          description: If false, fallback to cached content older than max-age if live fetch fails or times out. If true, returns an error instead.
          default: false
      type: object
      title: FetchPolicy
      description: Policy for live fetching web results.
    V1ExtractRequest:
      properties:
        urls:
          items:
            type: string
          type: array
          title: Urls
          description: URLs to extract content from. Up to 20 URLs.
        objective:
          anyOf:
          - type: string
          - type: 'null'
          title: Objective
          description: As in SearchRequest, a natural-language description of the underlying question or goal driving the request. Used together with search_queries to focus excerpts on the most relevant content.
        search_queries:
          anyOf:
          - items:
              type: string
            type: array
          - type: 'null'
          title: Search Queries
          description: Optional keyword search queries, as in SearchRequest. Used together with objective to focus excerpts on the most relevant content.
        max_chars_total:
          anyOf:
          - type: integer
          - type: 'null'
          title: Max Chars Total
          description: Upper bound on total characters across excerpts from all extracted results.
        session_id:
          anyOf:
          - type: string
            maxLength: 1000
          - type: 'null'
          title: Session Id
          description: Session identifier to track calls across separate search and extract calls, to be used as part of a larger task. Specifying it may give better contextual results for subsequent API calls.
        client_model:
          anyOf:
          - type: string
          - type: 'null'
          title: Client Model
          description: The model generating this request and consuming the results. Enables optimizations and tailors default settings for the model's capabilities.
          examples:
          - claude-opus-4-7
          - gpt-5.4
          - gemini-3.1-pro
        advanced_settings:
          anyOf:
          - $ref: '#/components/schemas/AdvancedExtractSettings'
          - type: 'null'
          description: Advanced configuration for fetch policy, excerpt settings, and full content settings. May impact result quality and latency unless used carefully. When omitted, excerpts are enabled and full content is disabled by default.
      additionalProperties: false
      type: object
      required:
      - urls
      title: V1ExtractRequest
      description: Extract request.
    V1ExtractResponse:
      properties:
        extract_id:
          type: string
          title: Extract Id
          description: Extract request ID, e.g. `extract_cad0a6d2dec046bd95ae900527d880e7`
        results:
          items:
            $ref: '#/components/schemas/V1ExtractResult'
          type: array
          title: Results
          description: Successful extract results.
        errors:
          items:
            $ref: '#/components/schemas/ExtractError'
          type: array
          title: Errors
          description: 'Extract errors: requested URLs not in the results.'
        warnings:
          anyOf:
          - items:
              $ref: '#/components/schemas/Warning'
            type: array
          - type: 'null'
          title: Warnings
          description: Warnings for the extract request, if any.
        usage:
          anyOf:
          - items:
              $ref: '#/components/schemas/UsageItem'
            type: array
          - type: 'null'
          title: Usage
          description: Usage metrics for the extract request.
        session_id:
          type: string
          title: Session Id
          description: Session identifier. Echoed back from the request if provided, otherwise generated by the server. Should be passed to future search and extract calls made by the agent as part of the same larger task.
          examples:
          - session_8a911eb27c7a4afaa20d0d9dc98d07c0
      type: object
      required:
      - extract_id
      - results
      - errors
      - session_id
      title: V1ExtractResponse
      description: Extract response.
    Error:
      properties:
        ref_id:
          type: string
          title: Reference ID
          description: Reference ID for the error.
        message:
          type: string
          title: Message
          description: Human-readable message.
        detail:
          anyOf:
          - additionalProperties: true
            type: object
          - type: 'null'
          title: Detail
          description: Optional detail supporting the error.
      type: object
      required:
      - ref_id
      - message
      title: Error
      description: An error message.
  securitySchemes:
    ApiKeyAuth:
      type: apiKey
      in: header
      name: x-api-key