SignalShelf · Developer documentation

instagram · media · transcript

Transcript of a public Instagram reel/video post (ASR). GET /v2/instagram/media/transcript accepts your x-api-key header. Costs credits that scale with the request (see parameters) per successful fresh call; failures and opt-in cache hits (cache_max_age) are free. Native path: GET /api/v1/web/instagram/media/transcript. Same request parameters and JSON shape as the reference API, so you can switch by changing the base URL.

Request parameters

Runs speech-to-text on the public media audio (audio-only download, whisper). Resolves the post via the guest instagram/post op; one transcript per video item (carousels up to 5). Posts without video return 404.

ParameterTypeRequiredDescription
urlstringYesPublic Instagram post or reel URL.
allow_long_transcriptionbooleanNoTranscribe up to 10 min (ASR_MAX_SECONDS) instead of the first 2 min (ASR_SHORT_SECONDS). Long transcription may cost extra credits.
languagestringNoSpoken language hint (ISO-639-1, e.g. "en"); default auto-detect.
cache_max_agestringNoServe a cached result up to this age (e.g. 1d, 7d). Handled by the warm store.
freshnessstringNo24h (default), 1h, 7d, or live. Accepted cache hits are free; live forces refresh except simultaneous requests may share newly fetched data.

Code examples

Set SIGNALSHELF_BASE_URL and SIGNALSHELF_API_KEY on your server.

REQUEST EXAMPLE
curl "$SIGNALSHELF_BASE_URL/v2/instagram/media/transcript?url=https%3A%2F%2Fwww.instagram.com%2Freel%2FDfJThjLOoPn%2F" \
  -H "x-api-key: $SIGNALSHELF_API_KEY"

Response and pagination

The response contains data and meta. data includes platform, resource, sample:false, provider, and the public source payload. Lists return data.items; individual resources return data.data.

Author feeds include data.next_cursor. Return it unchanged in cursor for the next page. Feed entries can include reposts. Single-post lookups resolve the handle to a DID first. Deleted or moderated posts may be unavailable.

meta includes event_id, mode, cache, credits_charged, fetched_at, age_seconds, and duration_ms. A fresh successful result costs one credit; accepted cached results and failures cost zero.

Availability and failures

Implementation is not a continuous availability guarantee. Source rate limits, retention, deleted content, restrictions, and regional availability apply. This is an independent SignalShelf contract, not a drop-in competitor response format.

400: invalid parameters. 401: missing/revoked key. 402: insufficient credits. 404: missing public resource. 429: SignalShelf throttle. 502/503: upstream failure or rate limit, with credit restored.

Explore the API in your workspace. Sample results and accepted cache hits cost zero credits. Open the playground →