framefetch-net Transcript API

openapi: 3.2.0
info:
  title: Framefetch Net Transcript API
  version: 1.7.0
  contact:
    name: FrameFetch
    url: https://framefetch.net
  termsOfService: https://framefetch.net/impressum
  description: 'Operations tagged Transcript across 2 of this provider''s published API definitions: openapi.json, framefetch-net-openapi.json. Each path carries the servers of the definition it was published in.'
servers:
- url: https://framefetch.net
security:
- bearerAuth: []
tags:
- name: Transcript
paths:
  /v1/transcript:
    post:
      operationId: getTranscript
      summary: Transcript only
      description: The "transcript" field is always forced in, so the optional top-level `translate` param (ISO-639-1 target language) and ?format=srt|vtt subtitle export both work here exactly as on /v1/extract.
      parameters:
      - $ref: '#/components/parameters/FormatMd'
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/UrlBody'
      responses:
        '200':
          description: Result
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ExtractResult'
        '409':
          description: SUBTITLES_UNAVAILABLE — ?format=srt|vtt was requested but the result carries no timed transcript segments.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
      tags:
      - Transcript
    servers:
    - url: https://framefetch.net
components:
  schemas:
    VideoComments:
      type: object
      description: 'Up to cap_applied TOP-LEVEL comments for a video, sorted by the platform''s "top" ranking (present only when "comments" was requested). Supported on YouTube only (via yt-dlp''s info-json comments). Reddit''s own public comment source — the unauthenticated thread JSON API — worked the same way until Reddit deprecated unauthenticated access to it on 2026-05-28; TikTok/Instagram/Pinterest never had a reliable public comment source. Billed a flat $0.0045 per call (independent of comments_cap), charged only when comments were actually produced. Best-effort: any fetch/parse failure, or an unsupported platform, omits the whole field with a warning and is NOT charged — never surfaces as an HTTP error code. Worker-internal failure codes behind those warnings (never an HTTP error, only a warnings[] string): COMMENTS_UNSUPPORTED (the platform has no comment source — drop "comments" from fields for it; see GET /v1/platforms), COMMENTS_PROVIDER_ERROR (the yt-dlp info-json fetch or parse failed).'
      properties:
        items:
          type: array
          items:
            $ref: '#/components/schemas/VideoComment'
        total_fetched:
          type: integer
          description: How many comments actually came back (after the cap).
        cap_applied:
          type: integer
          description: The effective cap that was applied (default 30, hard max 200 via comments_cap).
        sort:
          type: string
          enum:
          - top
    Error:
      type: object
      properties:
        error:
          type: object
          properties:
            code:
              type: string
            message:
              type: string
            hint:
              type: string
    CommentSentiment:
      type: object
      description: 'Aggregated LLM sentiment over a video''s comments (present only when "comment_sentiment" was requested, which auto-includes "comments" as its input): ONE Groq chat call (llama-3.3-70b-versatile, temperature 0, JSON mode) reads up to 100 comments and returns the audience mood AS A WHOLE — not per-comment scoring. Requires at least 5 fetched comments; fewer (or a model/parse failure) omits the whole field with a warning and is NOT charged — never surfaces as an HTTP error code. Billed a flat $0.006 per call, on top of the $0.0045 "comments" call it rides on. No competitor offers this rollup.'
      properties:
        positive_pct:
          type: number
          description: 0-100; positive_pct + neutral_pct + negative_pct === 100 (normalized).
        neutral_pct:
          type: number
        negative_pct:
          type: number
        summary:
          type: string
          description: 2-3 plain English sentences characterizing the overall audience reaction.
        top_themes:
          type: array
          items:
            type: string
          description: Up to 5 short recurring themes across the comments.
        representative:
          type: object
          description: One exemplar comment per polarity, when found.
          properties:
            positive:
              type: string
            negative:
              type: string
    TranscriptSegment:
      type: object
      description: 'One timed transcript segment: start/end offsets in seconds (floats, relative to the video start) and the segment''s text.'
      required:
      - start
      - end
      - text
      properties:
        start:
          type: number
        end:
          type: number
        text:
          type: string
    VideoStructured:
      type: object
      description: 'Typed, structured understanding of a video (present only when "structured" was requested): the transcript plus up to 12 sampled keyframes are run through a vision LLM (gpt-4o-mini, temperature 0) that returns chapters, entities, products shown, claims, and key moments as strict JSON. Deterministic and cacheable — unlike audio_digest it carries no expiring signed URL, so it takes part in the normal content-hash cache. Best-effort like digest: any download/VLM/parse failure omits the whole field, adds a warning to the top-level warnings array, and is NOT charged; the rest of the extraction stays usable. Worker-internal failure codes (never surfaced as an HTTP error, only as a warnings[] string): STRUCTURED_UNAVAILABLE (no vision provider configured — OPENAI_API_KEY unset), STRUCTURED_PROVIDER_ERROR (the provider returned a non-2xx response), STRUCTURED_PARSE_ERROR (the model''s output could not be parsed/validated into this schema). Billed a flat $0.03 per video (perStructuredVideoMicros, floor-pinned — see /pricing) charged only when the analysis is actually produced.'
      properties:
        chapters:
          type: array
          description: Timespan segments of the video.
          items:
            type: object
            properties:
              start_sec:
                type: number
              end_sec:
                type: number
              title:
                type: string
              summary:
                type: string
        entities:
          type: array
          description: People/orgs/places/products/other the model saw or heard.
          items:
            type: object
            properties:
              name:
                type: string
              type:
                type: string
                enum:
                - person
                - org
                - place
                - product
                - other
        products_shown:
          type: array
          description: Products/brands the video shows and/or talks about.
          items:
            type: object
            properties:
              name:
                type: string
              evidence:
                type: string
                enum:
                - visual
                - spoken
                - both
                description: 'How the product was surfaced: only on-screen, only mentioned, or both.'
        claims:
          type: array
          description: Factual claims made in the video, optionally anchored to a timestamp.
          items:
            type: object
            properties:
              text:
                type: string
              timestamp_sec:
                type: number
        key_moments:
          type: array
          description: Notable moments (hooks, reveals, demo steps) at a specific second.
          items:
            type: object
            properties:
              timestamp_sec:
                type: number
              description:
                type: string
        keyframe_mode:
          type: string
          enum:
          - scene
          - uniform
          description: 'How the analyzed frames were selected: ''scene'' = content-aware cut detection, ''uniform'' = even-sampling fallback for static/short clips with too few detected cuts.'
    ExtractResult:
      type: object
      properties:
        platform:
          type: string
        url:
          type: string
        metadata:
          type: object
          properties:
            title:
              type:
              - string
              - 'null'
            uploader:
              type:
              - string
              - 'null'
            durationSec:
              type:
              - number
              - 'null'
            uploadDate:
              type:
              - string
              - 'null'
              description: normalized YYYY-MM-DD
            sourceFps:
              type:
              - number
              - 'null'
            thumbnail:
              type:
              - string
              - 'null'
              description: poster/cover image URL (additive; may be null or omitted for older callers).
        insights:
          type: object
        transcript:
          type: object
          properties:
            text:
              type: string
            source:
              type: string
            lang:
              type: string
              description: Best-effort source language code, when known.
            segments:
              type: array
              description: 'Timed transcript segments (Whisper verbose_json or caption cue timings). Additive: omitted entirely (never []) when the source carries no usable timings.'
              items:
                $ref: '#/components/schemas/TranscriptSegment'
        transcript_translated:
          type: object
          description: Present only when the caller passed the `translate` param. An LLM translation of the transcript into the requested ISO-639-1 language, billed per audio-minute like transcription.
          properties:
            lang:
              type: string
              description: The requested ISO-639-1 target language.
            text:
              type: string
            segments:
              type: array
              description: Mirrors the source cue timings (start/end unchanged, text translated); omitted when the source transcript carried no timings — never back-filled from the untranslated transcript.
              items:
                $ref: '#/components/schemas/TranscriptSegment'
        digest:
          type: object
          description: LLM summary of the transcript (present only when the digest field was requested and a transcript was available).
          properties:
            gist:
              type: string
            topics:
              type: array
              items:
                type: string
            state:
              type: object
        audio_digest:
          $ref: '#/components/schemas/AudioDigest'
        structured:
          $ref: '#/components/schemas/VideoStructured'
        comments:
          $ref: '#/components/schemas/VideoComments'
        comment_sentiment:
          $ref: '#/components/schemas/CommentSentiment'
        delta:
          type: object
          description: 'Change since the most-recent prior fetch of the same url (present only when the delta field was requested). First time a url is seen -> { hasPrior: false } and nothing else. A per-hour rate is null when a counter was unavailable or no time had elapsed.'
          properties:
            hasPrior:
              type: boolean
            sinceTs:
              type: integer
              description: epoch ms of the prior snapshot this delta is measured against
            viewsPerHr:
              type:
              - number
              - 'null'
            likesPerHr:
              type:
              - number
              - 'null'
            commentsPerHr:
              type:
              - number
              - 'null'
            titleChanged:
              type: boolean
            thumbnailChanged:
              type: boolean
        frames:
          type: array
          items:
            type: object
            properties:
              index:
                type: integer
              url:
                type: string
              tSec:
                type: number
        textOverlay:
          type: array
          description: On-screen text per frame (OCR), same index as frames.
          items:
            type: object
            properties:
              index:
                type: integer
              text:
                type: string
              lines:
                type: array
                items:
                  type: object
                  properties:
                    text:
                      type: string
                    confidence:
                      type: number
                    bbox:
                      type: array
                      items:
                        type: number
                      minItems: 4
                      maxItems: 4
        cost:
          type: object
          properties:
            totalMicros:
              type: integer
            breakdownMicros:
              type: object
    VideoComment:
      type: object
      description: 'One top-level comment. Data-minimized by design: only the comment text, the author''s DISPLAY handle (never a user id, profile URL, or avatar), the like count, and the number of direct replies — all optional except text, omitted when the source doesn''t provide them.'
      required:
      - text
      properties:
        text:
          type: string
        author:
          type: string
          description: Display handle only — no user id / profile URL (data minimization).
        like_count:
          type: integer
        reply_count:
          type: integer
          description: Number of direct replies to this top-level comment.
    AudioDigest:
      type: object
      description: 'A spoken-audio rendering of the video''s digest (present only when "audio_digest" was requested and a digest was available to synthesize). Never served from the content-hash cache — its signed URL expires and the voice can vary per request. Billed per estimated spoken second (from the digest text length, ~15 chars/s) ONLY when the mp3 is actually produced; a TTS/upload failure omits this field, adds a warning to the top-level warnings array, and is not charged — the rest of the extraction stays usable. Failure reasons include no TTS provider configured (analogous to the worker-internal TTS_UNAVAILABLE code: neither OPENAI_API_KEY nor GROQ_API_KEY set) and an unsupported language on the only configured provider (TTS_LANG_UNSUPPORTED: Groq/PlayAI only speaks English/Arabic — set an OpenAI key for other languages); neither surfaces as an HTTP error code, only as a warnings[] string.'
      properties:
        url:
          type: string
          description: Signed URL to the mp3 (24h TTL).
        seconds:
          type: number
          description: Estimated spoken duration (digest gist chars / 15), the billing basis — not measured from the mp3.
        voice:
          type: string
          description: The TTS voice actually used (after per-provider fallback/mapping).
        provider:
          type: string
          enum:
          - openai
          - groq
        lang:
          type: string
          description: ISO-639-1 language the audio is spoken in; present only when a translate target drove a target-language digest.
    UrlBody:
      type: object
      required:
      - url
      properties:
        url:
          type: string
          format: uri
          description: 'Supported hosts only: youtube.com, youtu.be, tiktok.com, instagram.com, pinterest.*, reddit.com, redd.it'
  parameters:
    FormatMd:
      name: format
      in: query
      required: false
      schema:
        type: string
        enum:
        - md
        - markdown
        - srt
        - vtt
      description: 'Render the response in an alternate view instead of JSON. ''md''/''markdown'' returns a Markdown rendering (Content-Type: text/plain), equivalent to sending an Accept: text/plain or text/markdown header. ''srt''/''vtt'' renders the transcript''s timed segments as subtitles (Content-Type: application/x-subrip or text/vtt) — built from the TRANSLATED segments when the translate param was set, otherwise the source transcript''s segments; this is purely an egress view, so it reuses the same cached ExtractResult, cost, and cache key as the plain JSON call. Opt-in only (never inferred from an ambiguous Accept header); JSON stays the default. Extract-family endpoints only (/v1/extract, /v1/metadata, /v1/transcript, /v1/frames); srt/vtt additionally require "transcript" in fields (or "digest", which pulls it in) and return 409 SUBTITLES_UNAVAILABLE if the result has no timed segments (e.g. an older cached entry).'
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      description: 'API key from POST /v1/keys, sent as ''Authorization: Bearer <key>''.'
x-refined-from:
- openapi.json
- framefetch-net-openapi.json