Pay-per-success billing

The transcription API that only charges when it works

Real speech recognition — not caption scraping. Any audio or video file, any URL, meetings included. Speaker labels, timestamps, 99 languages. A failed job costs you nothing.

Create your API key Interactive API docs

Two requests to a transcript

Submit a file or URL, poll the job, done. Exports as JSON, SRT, WebVTT, or plain text.

1 · Submit

curl -X POST https://convertaudiototext.com/v1/transcribe \
  -H "Authorization: Bearer ck_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "source": "url",
    "input_url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
    "language": "auto"
  }'
# -> 201 { "job_id": "…", "status": "queued" }

2 · Fetch the result

curl https://convertaudiototext.com/v1/transcribe/JOB_ID \
  -H "Authorization: Bearer ck_live_YOUR_KEY"
# -> transcript with text, speakers, timestamps, confidence

# or grab it as subtitles:
curl https://convertaudiototext.com/v1/transcribe/JOB_ID/srt \
  -H "Authorization: Bearer ck_live_YOUR_KEY"

Full reference, schemas, and error codes in the interactive docs (OpenAPI 3.0, openapi.yaml). Agents: see auth.md and the API catalog.

Real ASR, not a caption reseller

Most “transcript APIs” re-serve YouTube's own captions. If a video has none, they fail — and none of them can tell you who spoke.

	ConvertAudioToText	Caption-scraper APIs
Videos without captions	Transcribed with real ASR	“Transcript could not be fetched”
Speaker labels (diarization)	Who said what, per segment	Not available
Sources	Files, URLs, podcasts, meetings	YouTube only
Punctuation & paragraphs	Full punctuation, readable text	Raw caption fragments
Failed requests	Never charged	Varies

Simple pricing, no surprise bills

Business

$59.99/month

Full API access with scoped, revocable keys
Webhooks for job completion
Speaker diarization on every transcript
All export formats (JSON, SRT, VTT, TXT)
Dedicated support

Get Business

Pay as you go

$9.99/100 minutes

Credit packs: 100, 500, or 1000 minutes
Credits never expire
Only successful transcripts consume credits
Larger packs cost less per minute

Buy minutes

Every plan is pay-per-success: a job that fails is never billed.

Questions developers ask

What does pay-per-success mean?

Minutes are deducted only when a transcription completes successfully. A job that fails — unreachable URL, restricted video, engine error — costs nothing.

Does it work on videos without captions?

Yes. The API runs real automatic speech recognition on the audio itself, so it transcribes any speech — it does not depend on YouTube captions existing.

Which sources are supported?

Direct file uploads (MP3, WAV, M4A, MP4, MOV and more), URLs (YouTube, Vimeo, SoundCloud, podcast feeds, any direct media link), and meeting recordings.

How do I authenticate?

Create an API key in the dashboard (Developers section) and send it as a Bearer token: Authorization: Bearer ck_live_... Keys are scoped and revocable.