Transcribe audio or video
/v1/transcribeUp to 60 minutes and 50 MB per request: MP3, WAV, M4A/AAC, FLAC, OGG/Opus, WebM, MP4 and
other common audio or video files. 25 European languages, detected automatically, with
punctuation and capitalisation. json returns the transcript with segment (and optionally
word) timings; srt and vtt return subtitles; txt the plain transcript.
Body parameters
formatjson · srt · vtt · txt`json` (transcript, segments, timings), `srt`, `vtt` or `txt`
"json"word_timestampsbooleanInclude every word with its start, end and confidence (json)
falsemax_charsintegerLongest segment or subtitle cue, in characters
84min 20max 200max_secondsnumberLongest segment or subtitle cue, in seconds
7min 1max 30audioobjectREQUIRED2 child fields
urlstringPublic http(s) URL of the audio or video file
2048base64stringThe file as base64 or a data: URI
formatjson · srt · vtt · txt`json` (transcript, segments, timings), `srt`, `vtt` or `txt`
"json"word_timestampsbooleanInclude every word with its start, end and confidence (json)
falsemax_charsintegerLongest segment or subtitle cue, in characters
84min 20max 200max_secondsnumberLongest segment or subtitle cue, in seconds
7min 1max 30filebinaryREQUIREDResponses
The transcript (json) or subtitles (srt, vtt, txt)
textstringREQUIREDThe whole transcript, punctuated and capitalised
durationnumberREQUIREDAudio duration in seconds
billed_minutesintegerREQUIREDAudio minutes this request counts against the plan (rounded up, at least 1)
confidencenumberREQUIREDAverage token confidence, 0–1
segmentsobject[]REQUIREDItem fields
startnumberREQUIREDendnumberREQUIREDtextstringREQUIREDwordsobject[]Present with `word_timestamps`
Item fields
wordstringREQUIREDstartnumberREQUIREDSeconds
endnumberREQUIREDconfidencenumberREQUIRED0–1
Binary response body.
Binary response body.
Binary response body.