Transcription API · MCP server

Transcription API,
minus the pipeline.

Three REST calls or one MCP tool call. Speaker-labelled JSON, SRT, VTT, TXT and PDF back.

Your first transcription is free · then 5 minutes a month · no card

REST APIMCP for AI agents

live API# connect once (Claude Code; Cursor and VS Code take the same URL)
claude mcp add --transport http -s user audiotranscribr \
  https://api.audiotranscribr.com/v1/mcp --header "x-api-key: $KEY"

# then just ask
> Transcribe https://audiotranscribr.com/samples/clip-22-451-01.mp3
  and give me the action items by speaker, plus the PDF.

# the agent calls transcribe_url and gets back
# the text, speaker labels and five file links

# Claude.ai: add https://api.audiotranscribr.com/mcp as a connector
# and sign in (no key)

live API# 1. create a transcription
curl -X POST https://api.audiotranscribr.com/v1/transcriptions \
  -H "x-api-key: $KEY" \
  -d '{"url": "https://audiotranscribr.com/samples/clip-22-451-01.mp3"}'

# 2. poll until it is done
{
  "status": "completed",
  "duration_seconds": 312.6,
  "language": "en",
  "results": {
    "json": "https://…", "srt": "https://…",
    "vtt": "https://…", "txt": "https://…",
    "pdf": "https://…"
  }
}

Built for AI agents, too

The API is also an MCP server. Connect Claude Code, Cursor or VS Code to one URL and your agent can transcribe a recording, read it, and hand back subtitles and a PDF — in a single tool call.

Set up the MCP server

Same API key. Same plans. Nothing to install. In Claude.ai, add https://api.audiotranscribr.com/mcp as a connector and sign in with your email: no key needed.

# Claude Code
claude mcp add --transport http -s user audiotranscribr \
  https://api.audiotranscribr.com/v1/mcp \
  --header "x-api-key: $KEY"

# then just ask
> Transcribe https://audiotranscribr.com/samples/clip-22-451-01.mp3
  and list the action items by speaker.

Three calls. That’s the integration.

One header for auth. No SDK, no webhooks to stand up, no queue to run.

Create

POST /v1/transcriptions
{ "url": "https://…/call.mp3" }

# or omit url and PUT the file
# to the upload link we return

Poll

GET /v1/transcriptions/{id}

{ "status": "transcribing" }
{ "status": "completed", … }

Download

GET /v1/transcriptions/{id}
      /result?format=pdf

# json · srt · vtt · txt · pdf

Full reference with a live “try it” console →

Use it from your tools

Pick where you work. Every path uses the same account, plans and files.

DashboardNo code: upload a file or paste a YouTube link, download the files.Claude.ai and ChatGPTAdd the MCP connector, sign in with your email, ask for a transcript in chat.Claude Code, Cursor, VS CodeOne URL and your API key; the agent transcribes, reads and fetches files.REST APIThree calls: create, poll, download. curl, JavaScript, Python and PHP examples.n8nHTTP Request, Wait, HTTP Request, IF. Workflow file to import.MakeHTTP module, Sleep, HTTP module, a filter. Nothing to install.ZapierWebhooks by Zapier, Delay, Webhooks, Filter.PostmanPublic workspace with every request ready to fork.OpenAPIGenerate a client in any language from the spec.

Guides

Working recipes for common jobs.

Pricing

Your first transcription is free, then 5 minutes a month. Paid plans start at $9 a month for 10 hours of audio, with per-second overage; pay-as-you-go is on the plans page. All plans →

Transcribe your first file in five minutes

Create an account, make a key, paste the curl command, download a PDF.

Transcription API,minus the pipeline.