Screenpipe Frames API

openapi: 3.2.0
info:
  title: Screenpipe Frames API
  version: 1.0.0
  description: 'Screenpipe captures everything you see, say, and hear on your computer. Use this API to search through captured content, manage recordings, and build AI-powered automations on top of your screen data.


    The server runs locally at `http://localhost:3030` by default.'
tags:
- name: Frames
  description: Access captured screenshots and their extracted text
paths:
  /frames/{frame_id}:
    get:
      operationId: routes_frames_get_frame_data
      parameters:
      - name: frame_id
        schema:
          type: integer
        in: path
        required: true
        style: simple
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema: {}
      tags:
      - Frames
      summary: Get frame by ID
      description: Returns a captured screenshot frame with optional base64 image data.
  /frames/{frame_id}/text:
    get:
      description: 'Get frame text positions with bounding boxes for a specific frame.

        Uses accessibility tree node bounds when available, and OCR fallback positions for visual-only or legacy frames.

        Both OCR and accessibility bounds are normalized to 0-1 relative to the

        monitor (full-screen capture), so they align correctly with the screenshot.'
      operationId: routes_frames_get_frame_text_data
      parameters:
      - name: frame_id
        schema:
          type: integer
        in: path
        required: true
        style: simple
      - name: query
        schema:
          type:
          - string
          - 'null'
        in: query
        style: form
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/FrameTextResponse'
      tags:
      - Frames
      summary: Get frame text and bounds
    post:
      description: 'Run on-demand OCR on a frame that has no stored bounding boxes.

        Loads the snapshot JPEG, runs Apple Vision OCR, stores the result,

        and returns the text positions. Subsequent GET requests will hit the

        cached DB row. If OCR data already exists, returns it without re-running.'
      operationId: routes_frames_run_frame_ocr
      parameters:
      - name: frame_id
        schema:
          type: integer
        in: path
        required: true
        style: simple
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/FrameTextResponse'
      tags:
      - Frames
      summary: Run OCR on frame
  /frames/{frame_id}/ocr:
    get:
      description: 'Get frame text positions with bounding boxes for a specific frame.

        Uses accessibility tree node bounds when available, and OCR fallback positions for visual-only or legacy frames.

        Both OCR and accessibility bounds are normalized to 0-1 relative to the

        monitor (full-screen capture), so they align correctly with the screenshot.'
      operationId: getFramesByFrameIdOcr
      parameters:
      - name: frame_id
        schema:
          type: integer
        in: path
        required: true
        style: simple
      - name: query
        schema:
          type:
          - string
          - 'null'
        in: query
        style: form
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/FrameTextResponse'
      tags:
      - Frames
      summary: Get frame OCR (deprecated)
      x-operation-id-source: normalized
      x-operation-id-original: routes_frames_get_frame_text_data
    post:
      description: 'Run on-demand OCR on a frame that has no stored bounding boxes.

        Loads the snapshot JPEG, runs Apple Vision OCR, stores the result,

        and returns the text positions. Subsequent GET requests will hit the

        cached DB row. If OCR data already exists, returns it without re-running.'
      operationId: postFramesByFrameIdOcr
      parameters:
      - name: frame_id
        schema:
          type: integer
        in: path
        required: true
        style: simple
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/FrameTextResponse'
      tags:
      - Frames
      summary: Run frame OCR (deprecated)
      x-operation-id-source: normalized
      x-operation-id-original: routes_frames_run_frame_ocr
  /frames/{frame_id}/context:
    get:
      description: 'Get frame context: accessibility text, tree nodes, and extracted URLs.

        Falls back to OCR data for legacy frames without accessibility data.'
      operationId: routes_frames_get_frame_context
      parameters:
      - name: frame_id
        schema:
          type: integer
        in: path
        required: true
        style: simple
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/FrameContextResponse'
      tags:
      - Frames
      summary: Get frame context
  /frames/{frame_id}/metadata:
    get:
      description: Get frame metadata (timestamp) for deep link navigation. screenpipe://frame/123 → resolve to timestamp.
      operationId: routes_frames_get_frame_metadata
      parameters:
      - name: frame_id
        schema:
          type: integer
        in: path
        required: true
        style: simple
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/FrameMetadataResponse'
      tags:
      - Frames
      summary: Get frame metadata
  /frames/next-valid:
    get:
      description: 'Find the next frame that has a valid video file on disk.

        This allows the frontend to skip directly to a valid frame instead of

        trying each frame one-by-one when frames fail to load.'
      operationId: routes_frames_get_next_valid_frame
      parameters:
      - name: frame_id
        schema:
          type: integer
        in: query
        style: form
      - name: direction
        schema:
          type: string
        in: query
        style: form
      - name: limit
        schema:
          type: integer
        in: query
        style: form
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/NextValidFrameResponse'
      tags:
      - Frames
      summary: Get next valid frame
  /frames/{frame_id}/elements:
    get:
      description: Get all elements for a specific frame (full element tree).
      operationId: routes_elements_get_frame_elements
      parameters:
      - name: frame_id
        schema:
          type: integer
        in: path
        required: true
        style: simple
      - name: source
        schema:
          type:
          - string
          - 'null'
        in: query
        style: form
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ElementsListResponse'
      tags:
      - Frames
      summary: Get frame UI elements
components:
  schemas:
    FrameTextResponse:
      description: Response type for frame OCR data endpoint
      type: object
      properties:
        frame_id:
          type: integer
        text_positions:
          type: array
          items:
            $ref: '#/components/schemas/TextPosition'
      required:
      - frame_id
      - text_positions
    AccessibilityNodeBounds:
      description: Bounding box for an accessibility node (0-1 normalized to monitor)
      type: object
      properties:
        left:
          type: number
        top:
          type: number
        width:
          type: number
        height:
          type: number
      required:
      - left
      - top
      - width
      - height
    TextPosition:
      type: object
      properties:
        text:
          type: string
        confidence:
          type: number
        bounds:
          $ref: '#/components/schemas/TextBounds'
      required:
      - text
      - confidence
      - bounds
    FrameContextResponse:
      description: Response type for frame context endpoint (accessibility-first, OCR fallback)
      type: object
      properties:
        frame_id:
          type: integer
        text:
          type:
          - string
          - 'null'
        nodes:
          type: array
          items:
            $ref: '#/components/schemas/AccessibilityNode'
        urls:
          type: array
          items:
            type: string
        text_source:
          type: string
      required:
      - frame_id
      - text
      - nodes
      - urls
      - text_source
    AccessibilityNode:
      description: A node from the accessibility tree
      type: object
      properties:
        role:
          type: string
        text:
          type: string
        depth:
          type: integer
        bounds:
          $ref: '#/components/schemas/AccessibilityNodeBounds'
        properties:
          type:
          - object
          - 'null'
      required:
      - role
      - text
      - depth
    ElementsListResponse:
      type: object
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/ElementResponse'
        pagination:
          $ref: '#/components/schemas/PaginationResponse'
      required:
      - data
      - pagination
    BoundsResponse:
      type: object
      properties:
        left:
          type: number
        top:
          type: number
        width:
          type: number
        height:
          type: number
      required:
      - left
      - top
      - width
      - height
    FrameMetadataResponse:
      description: Response type for frame metadata (used by deep link navigation)
      type: object
      properties:
        frame_id:
          type: integer
        timestamp:
          type: string
          format: date-time
      required:
      - frame_id
      - timestamp
    NextValidFrameResponse:
      description: Response for next valid frame endpoint
      type: object
      properties:
        frame_id:
          type: integer
        timestamp:
          type: string
          format: date-time
        skipped_count:
          type: integer
      required:
      - frame_id
      - timestamp
      - skipped_count
    PaginationResponse:
      type: object
      properties:
        limit:
          type: integer
        offset:
          type: integer
        total:
          type: integer
      required:
      - limit
      - offset
      - total
    ElementResponse:
      type: object
      properties:
        id:
          type: integer
        frame_id:
          type: integer
        source:
          type: string
        role:
          type: string
        text:
          type:
          - string
          - 'null'
        parent_id:
          type:
          - integer
          - 'null'
        depth:
          type: integer
        bounds:
          $ref: '#/components/schemas/BoundsResponse'
        confidence:
          type:
          - number
          - 'null'
        sort_order:
          type: integer
      required:
      - id
      - frame_id
      - source
      - role
      - text
      - parent_id
      - depth
      - bounds
      - confidence
      - sort_order
    TextBounds:
      type: object
      properties:
        left:
          type: number
        top:
          type: number
        width:
          type: number
        height:
          type: number
      required:
      - left
      - top
      - width
      - height
x-provenance:
  method: harvested
  publisher: Screenpipe
  source: https://docs.screenpipe.com/openapi.yaml
  note: split by refine-openapis from the provider's own published document, verified by fetch (openapi/_original/*.fetch.json);
    stamped by stamp-openapi-provenance.py (roadmap#793)