Public API

What is a Speech-to-Text API

A Speech-to-Text API turns audio and video into text programmatically—without manual uploads in the browser. The Polyglot Voice API fits bots, CRMs, call centers, edtech, and media products: send a file or URL, poll status or use a webhook, then fetch the transcript, translation, and export files (SRT, TXT, JSON). Processing runs in the Polyglot Voice cloud—you do not deploy or download recognition models on your side.

API capabilities

• Speech recognition to text (98+ languages)
• Transcript and video content translation
• Automatic subtitles and SRT/TXT export
• Interview, lecture, and call transcription
• Webhooks on completion and status polling

Fast integration path

1. Step 1Create API key
2. Step 2POST upload/from-url
3. Step 3Receive webhook or poll status
4. Step 4GET result/download

Base URL

Use the base URL for all API requests.

https://polyglotvoice.ru/api/v1

Authorization

The endpoints below accept Authorization: Bearer with your site session JWT or an API key sk_….

Authorization: Bearer sk_...

The sk_… key is created in your profile, lasts 30 days, and applies only to transcriptions/*; the rest of the site uses JWT. Either credential works for these routes.

How to get an API key

1. Sign in to your account.
2. Open Profile → “API keys”.
3. Create a key and save it (shown once).
4. Use the key in requests.

Endpoints

Core endpoints for working with transcriptions.

Request examples

Basic examples to get started.

curl -X POST "https://polyglotvoice.ru/api/v1/transcriptions/upload" \
  -H "Authorization: Bearer sk_..." \
  -F "file=@audio.mp3"

import requests

url = "https://polyglotvoice.ru/api/v1/transcriptions/upload"
headers = {"Authorization": "Bearer sk_..."}
files = {"file": open("audio.mp3", "rb")}

response = requests.post(url, headers=headers, files=files)
print(response.json())

const formData = new FormData();
formData.append("file", file);

fetch("https://polyglotvoice.ru/api/v1/transcriptions/upload", {
  method: "POST",
  headers: { Authorization: "Bearer sk_..." },
  body: formData
})
  .then((res) => res.json())
  .then(console.log);

File formats

Popular audio and video formats are supported.

For example: mp3, wav, m4a, ogg, mp4, webm.

Task statuses

A transcription task moves through these statuses:

• queued — accepted, waiting to process
• processing — media is being processed
• completed — result ready; fetch text and downloads
• failed — processing error (see failure_reason)
• cancelled — cancelled by the user

Responses and errors

Example of task creation response and auth error.

{ "task_id": "9b9c1b9a-....", "status": "queued" }

{
  "task_id": "9b9c1b9a-....",
  "status": "completed",
  "model": "fast",
  "input_language": "en",
  "output_language": "ru",
  "audio_duration_seconds": 125.4,
  "confidence": 0.94,
  "transcript_preview": "Hello, this is a sample transcript..."
}

{
  "task_id": "9b9c1b9a-....",
  "transcript": "Hello, this is a sample transcript...",
  "translation": "Привет, это пример транскрипта...",
  "segments": [{ "start": 0.0, "end": 2.4, "text": "Hello..." }],
  "confidence": 0.94,
  "updated_at": "2026-05-24T12:00:00Z"
}

{ "detail": "Invalid token." }

Webhook

Set a webhook URL in your profile for the sk_ API key. When a task reaches a final status, Polyglot Voice sends an HTTP POST to your URL. Headers: Content-Type: application/json, X-Polyglot-Timestamp, X-Polyglot-Signature (HMAC-SHA256 over the body). Fetch the full transcript via GET /transcriptions/{task_id}/result.

Sample webhook body

POST https://client.example.com/webhook

{
  "event": "transcription.completed",
  "task_id": "9b9c1b9a-....",
  "status": "completed",
  "model": "fast",
  "failure_reason": null,
  "has_transcript": true,
  "has_translation": true,
  "transcript_preview": "Hello, this is a sample..."
}

HTTP error codes

Typical API error responses:

• 400 — bad request (parameters, format, URL)
• 401 — invalid or missing key/token
• 403 — access denied
• 404 — task or file not found
• 429 — rate limit exceeded (see Limits)
• 500 — internal server error

Limits

These values are enforced by the API. Short-term (burst) limits apply per signed-in user — the same account whether you use a session JWT or an sk_ key. Daily and monthly counters count every request made with any of your sk_ keys (UTC calendar day and month). When a limit is exceeded, the API responds with HTTP 429.

The live JSON field rate_limits is returned by:

Loading current limits…

In addition: each plan limits max audio length per file, max upload size, concurrent tasks, and (on the website) daily uploads. See your subscription on the dashboard.

FAQ

Related pages