ConvertAudioToText API reference

Send an audio or video file or a link, get a transcript back. Every endpoint, parameter and response is listed below; you can try calls with your API key. Plans and keys: convertaudiototext.com/transcription-api.

What the API does

Public transcription API for ConvertAudioToText. Submit an audio or video file (by upload or URL), poll for the result, and export the transcript as SRT, WebVTT, plain text, or JSON. Billing is pay-per-success: minutes are deducted only when a transcription completes successfully, and failed jobs are never charged. This document describes only the public developer surface served at /v1. The dashboard and admin APIs are not part of this contract and are not documented here.

Authentication

API key with the ck_live_ (production) or ck_test_ (test) prefix, sent as Authorization: Bearer ck_live_.... Create a key on the Developer or Business plan at https://convertaudiototext.com/dashboard/developers. Session JWTs are rejected on this API; only ck_live_/ck_test_ keys are accepted, and each must carry the transcribe scope.

Base URLs

Endpoints

POST /v1/transcribe: Create a transcription job

Submit a media file for transcription, either by public URL or by file upload. Returns immediately with a job id; the transcript is produced asynchronously. Requires an API key with the transcribe scope.

To upload a file, send multipart/form-data with source=upload and the media in the file field:

curl -X POST https://api.convertaudiototext.com/v1/transcribe -H "Authorization: Bearer ck_live_..." -F "source=upload" -F "file=@meeting.mp3" -F "language=en"

Accepted audio: mp3, wav, flac, m4a, ogg, opus, aac, wma, amr, aiff, aif, oga, caf, weba. Accepted video: mp4, mov, avi, mkv, webm, wmv, flv, 3gp, 3gpp, mpg, mpeg, m4v, mts, ts, m2ts, vob, mxf, ogv. The format is read from the file name's extension, and the file must contain an audio track. API plans accept files up to 10 GB, but one upload request is limited to 100 MB at our edge. For a larger file, put it at a URL the API can fetch (a public link or a presigned storage URL) and send it with source=url.

Responses:

GET /v1/transcribe: List transcription jobs

List the jobs on the account behind the API key, newest first. The response is one page: total is the number of jobs on the account, not the number returned, so page with limit and offset until has_more is false.

Parameters:

Responses:

GET /v1/transcribe/{job_id}: Get a transcription result

Read one job and its transcript. Poll this endpoint until status is completed.

Responses:

GET /v1/transcribe/{job_id}/json: Export as JSON

Responses:

GET /v1/transcribe/{job_id}/srt: Export as SRT subtitles

Responses:

GET /v1/transcribe/{job_id}/txt: Export as plain text

Responses:

GET /v1/transcribe/{job_id}/vtt: Export as WebVTT subtitles

Responses:

Full specification

The complete OpenAPI 3 file, with every schema and example, is at /openapi.yaml. The interactive reference below is built from the same file. Plans, prices and API keys: convertaudiototext.com/transcription-api.