naftiko: 1.0.0-alpha2
info:
  label: Bright Data Web Scraper
  description: |
    Asynchronous snapshot-based scraping over 660+ pre-built dataset endpoints. Trigger jobs, poll
    progress, list snapshots, cancel, rerun, and download results.
  tags:
  - Bright Data
  - Web Scraping
  - Datasets
  created: '2026-05-25'
  modified: '2026-05-25'
binds:
- namespace: env
  keys:
    BRIGHTDATA_API_TOKEN: BRIGHTDATA_API_TOKEN
capability:
  consumes:
  - type: http
    namespace: web-scraper
    baseUri: https://api.brightdata.com
    description: Bright Data Web Scraper API.
    resources:
    - name: scrape
      path: /datasets/v3/scrape
      operations:
      - name: triggerScrape
        method: POST
        description: Trigger an async scrape against a dataset_id.
        outputRawFormat: json
        outputParameters:
        - name: snapshot_id
          type: string
          value: $.snapshot_id
        inputParameters:
        - name: dataset_id
          in: query
          type: string
          required: true
        - name: body
          in: body
          type: array
          required: true
    - name: progress
      path: /datasets/v3/progress/{snapshot_id}
      operations:
      - name: getProgress
        method: GET
        description: Poll scrape progress.
        outputRawFormat: json
        outputParameters:
        - name: status
          type: string
          value: $.status
        inputParameters:
        - name: snapshot_id
          in: path
          type: string
          required: true
    - name: snapshot
      path: /datasets/v3/snapshot/{snapshot_id}
      operations:
      - name: downloadSnapshot
        method: GET
        description: Download snapshot results in JSON/NDJSON/CSV/Parquet.
        outputRawFormat: json
        outputParameters:
        - name: rows
          type: array
          value: $.
        inputParameters:
        - name: snapshot_id
          in: path
          type: string
          required: true
        - name: format
          in: query
          type: string
      - name: cancelSnapshot
        method: POST
        description: Cancel a running snapshot.
        path: /datasets/v3/snapshot/{snapshot_id}/cancel
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
        inputParameters:
        - name: snapshot_id
          in: path
          type: string
          required: true
    authentication:
      type: apikey
      key: Authorization
      value: 'Bearer {{env.BRIGHTDATA_API_TOKEN}}'
      placement: header
  exposes:
  - type: mcp
    namespace: web-scraper-mcp
    port: 9090
    transport: http
    description: MCP adapter for Bright Data Web Scraper.
    tools:
    - name: bright-data-scrape-trigger
      description: Trigger a scraping job against a Bright Data dataset.
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: web-scraper.triggerScrape
      with:
        dataset_id: tools.dataset_id
        body: tools.body
      outputParameters:
      - type: object
        mapping: $.
    - name: bright-data-scrape-progress
      description: Get progress of an async snapshot.
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: web-scraper.getProgress
      with:
        snapshot_id: tools.snapshot_id
      outputParameters:
      - type: object
        mapping: $.
    - name: bright-data-scrape-download
      description: Download a completed snapshot.
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: web-scraper.downloadSnapshot
      with:
        snapshot_id: tools.snapshot_id
        format: tools.format
      outputParameters:
      - type: object
        mapping: $.