# Speech to Text & Diarization

Speech to text (Whisper) with timestamps, SRT and VTT: clips up to 5 minutes in the same call, 3 hours queued, speaker diarization on queued jobs.

Send a direct link to an audio or video file, or the file itself. You get the transcript with timestamps as JSON, SRT and VTT. Clips up to 5 minutes are answered in the same call; longer recordings, up to three hours, are queued and usually finish within minutes, with a 24-hour promise: if we miss it, it's free. Pay per second of audio. On the queued call, set speakers to true to label who said each segment (S1, S2, ...).

- Long recordings are accepted: up to three hours per job, with JSON, SRT and VTT in one result.
- Speaker labels (S1, S2, ...) on queued jobs, at the same price.
- Pay per second of audio. A clip of up to 5 minutes is answered in the same call.

## Calls

### `POST https://readymadeapi.com/v1/transcript/instant`

Transcribe a short clip, up to 5 minutes and 15 MB, and get text, timestamps, SRT and VTT in the same call. max_seconds is the most the clip may last; you are charged only for the seconds it actually runs.

- Price: $0.000132 per second, up to 300 per call
- MCP tool: `transcript_intel_transcribe_now`
- Answers in the same call.

Request:

```json
{
  "audio_url": "https://readymadeapi.com/samples/voice-note.mp3",
  "max_seconds": 60
}
```

Answer:

```json
{
  "text": "Hi, this is a short voice note. Please move our meeting to Thursday at 10 and bring the signed contract.",
  "language": "en",
  "duration_s": 6.525,
  "segments": [
    {
      "start": 0.0,
      "end": 5.98,
      "text": "Hi, this is a short voice note. Please move our meeting to Thursday at 10 and bring the signed contract."
    }
  ],
  "srt": "1\n00:00:00,000 --> 00:00:05,980\nHi, this is a short voice note. Please move our meeting to Thursday at 10 and bring the signed contract.\n\n",
  "vtt": "WEBVTT\n\n00:00:00.000 --> 00:00:05.980\nHi, this is a short voice note. Please move our meeting to Thursday at 10 and bring the signed contract.\n\n"
}
```

### `POST https://readymadeapi.com/v1/transcript/transcribe`

Queue a transcription. max_minutes is the most the audio may last; you are charged only for the seconds it actually runs.

- Price: $0.000132 per second, up to 10800 per call
- MCP tool: `transcript_intel_transcribe`
- Queued: answers 202 with a job; poll `GET /v1/jobs/{id}`, fetch `GET /v1/jobs/{id}/result`. Charged up front, and what does not run is refunded.

Request:

```json
{
  "audio_url": "https://readymadeapi.com/samples/voice-note.mp3",
  "max_minutes": 60
}
```

The finished job:

```json
{
  "job_id": "01a1f553-8b2b-7487-b36c-93e3bd7be737",
  "status": "succeeded",
  "late": false,
  "deadline": "2026-11-20T19:40:55.463853Z",
  "completed_at": "2026-11-19T19:40:56.003566Z",
  "error": null,
  "result": {
    "text": "Hi, this is a short voice note. Please move our meeting to Thursday at 10 and bring the signed contract.",
    "language": "en",
    "segments": [
      {
        "end": 5.98,
        "text": "Hi, this is a short voice note. Please move our meeting to Thursday at 10 and bring the signed contract.",
        "start": 0.0
      }
    ],
    "duration_s": 6.525,
    "srt": "1\n00:00:00,000 --> 00:00:05,980\nHi, this is a short voice note. Please move our meeting to Thursday at 10 and bring the signed contract.\n\n",
    "vtt": "WEBVTT\n\n00:00:00.000 --> 00:00:05.980\nHi, this is a short voice note. Please move our meeting to Thursday at 10 and bring the signed contract.\n\n"
  }
}
```

## Limits

- Instant call: clips up to 5 minutes and 15 MB.
- Queued call: up to 3 hours of audio; by link, files up to 350 MB.
- Speaker labels: queued call only.

## Questions

**What is max_minutes for?**

It is your ceiling. Audio longer than that is refused and not charged; shorter audio is charged for its real length.

**Is the answer instant or queued?**

Both. /v1/transcript/instant answers in the same call for clips up to 5 minutes and 15 MB. /v1/transcript/transcribe queues longer files; most finish within minutes, the promise is 24 hours, and a missed deadline is free.

**Which languages?**

The 99 languages of Whisper large-v3-turbo. The language is detected, or you can name it. Accuracy varies by language; we test in English.

**Can I send a YouTube link?**

No. Send a direct link to an audio or video file, or the file itself.

**How is it billed?**

Transcription is billed per second of audio measured, not per request.

## Using it

- Send `Authorization: Bearer <key>` and a JSON body. Keys and credit: https://readymadeapi.com/account
- Every answer carries `X-Request-Id`, `X-Units` and `X-Charged-Usd`.
- An error is a JSON problem with a `title` to act on and a `detail` in plain words; nothing is charged for it.
- Full schema: https://readymadeapi.com/openapi.json
