Retrieve the transcript organized into individual sentences with start and end timestamps. Only available for completed transcriptions.
Returns the transcript text split into individual sentences, each with timing information (start and end timestamps in seconds). This provides finer granularity than the paragraphs endpoint. Supports pagination via limit and offset query parameters.
Supports ETag caching -- completed transcription sentences are immutable, so the API returns Cache-Control: public, max-age=86400. Use the If-None-Match header to avoid re-downloading unchanged data.
When PII redaction was enabled for the transcription, the response includes X-PII-Redacted and X-PII-Policies headers.
transcription_id string required
The transcription ID with trans_ prefix.
limit integer optional
Maximum number of sentences to return. Must be between 1 and 1000. Omit to return all sentences.
offset integer optional · Defaults to 0
Number of sentences to skip for pagination.
Returns a success envelope with:
data.transcription_id-- The transcription ID.data.audio_duration_seconds-- Total audio duration in seconds.data.sentences-- Array of sentence objects, each containingtext,start(seconds), andend(seconds).metadata-- Request metadata includingrequest_id,timestamp,processing_time_ms,upstream_latency_ms,retry_attempts,cache_hit,total_count, andreturned_count.actions-- Links to related format endpoints (get_paragraphs,get_subtitles_srt, etc.) and pagination links (next,previous) when applicable.
Response headers include ETag, Last-Modified, and Cache-Control.
- Bash
- Python
- Python
- JavaScript
curl -s -X GET "https://api.aitronos.com/v1/audio/transcribe/trans_a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6/sentences?limit=20&offset=0" \
-H "Authorization: Bearer $ACCESS_TOKEN" | python3 -m json.tool