Skip to main content
Whisper Web

Whisper Web Developer API

Whisper API for audio and video transcription

Build with managed Whisper speech-to-text. Submit audio or video in 99 languages and get timestamped JSON, text and subtitles.

Get your API key Read the quickstart
$0.01 / minute · prepaid API pricing

POST /v1/transcriptions

curl https://whisperweb.dev/v1/transcriptions \
  -H "Authorization: Bearer $WHISPERWEB_API_KEY" \
  -H "Idempotency-Key: $REQUEST_ID" \
  -H "Content-Type: application/json" \
  -d '{
    "source": {"audio_url": "https://example.com/audio.mp3"},
    "task": "transcribe", "language": "auto"
  }'

Swipe the code to view long lines.

Set your server-side API key and a unique REQUEST_ID for each new task. Keep the same ID and body when retrying. Follow the quickstart to poll and download the result.

Languages
99

Automatic detection or a source-language code.

Maximum file size
5 GB

Decimal gigabytes, per audio or video file.

Maximum duration
10 hours

Per recording, checked before transcription.

API rate
$0.01/min

Transcription and English translation, billed by the second.

Built for integration

Whisper API capabilities

Connect recordings to your application with a managed speech-to-text workflow, structured results and explicit usage controls.

Explore the API reference

  • Speech-to-text in 99 languages

    Detect the spoken language automatically or specify a language code. Keep the original language in your transcript.

  • Private uploads or media URLs

    Upload audio and video directly to private storage, or submit a public media URL. Process files up to 5 GB and 10 hours.

  • Timestamped text and subtitles

    Retrieve text with segment start and end times in milliseconds. Use structured JSON, plain TXT, or SRT and VTT captions.

  • Audio-to-English translation

    Choose translate_to_english to turn spoken audio into English text, using the same $0.01-per-minute rate and task workflow.

  • Async jobs and signed webhooks

    Receive a task ID, follow validation and processing, and fetch the result. Use polling or signed completion and failure events.

  • Use it from AI agents (MCP)

    Connect Claude, ChatGPT, Claude Code, Codex or Cursor to https://whisperweb.dev/mcp, sign in, and transcribe files, media URLs or YouTube links from chat with the same API credits.

  • Scoped keys and spending controls

    Set key permissions and expiry, revoke access, and cap the credits authorized for a task. Track API usage in your Developer Console.

Request to result

How the transcription API works

Open the Playground
  1. 1

    Submit your recording

    Create a task with a direct media URL or the ID of a completed private upload. Set the source language and transcription task.

    POST /v1/transcriptions
  2. 2

    Follow the task

    Save the returned task ID. Poll for status at least five seconds apart, or receive a signed webhook when processing finishes.

    GET /v1/transcriptions/{id}
  3. 3

    Retrieve text or subtitles

    After completion, get JSON for your application or download TXT, SRT and VTT. Review the actual charge in your API usage history.

    GET /v1/transcriptions/{id}/result

202 Accepted · response excerpt

{
  "id": "tr_example",
  "status": "validating",
  "retry_after_ms": 5000
}

Swipe the code to view long lines.

A task ID confirms acceptance. Follow the status until processing completes, then retrieve the transcript or subtitle file.

cURL, Python and Node.js examples Webhook guide

Text your application can use

Transcript and subtitle formats

Use structured transcript data in your code, plain text in your content workflow, or timed captions in a video player.

See result schemas and downloads
JSON
Transcript text, source and output languages, media duration, and timestamped segments for your application.
TXT
Plain transcript text for documents, editing and downstream text processing.
SRT
SubRip subtitle cues with timestamps for video editors and publishing tools.
VTT
WebVTT captions for web video players and caption workflows.

Build speech-to-text into your workflow

Podcast and media production

Convert recorded episodes into text for editing and subtitle files for your publishing workflow.

Interview and research tools

Add interview transcripts to your application and use segment timestamps to reference the original recording.

Courses and recorded lectures

Create readable transcripts and SRT or VTT captions from recorded lessons in the spoken language.

Search and automation workflows

Feed transcript text into your own indexing or analysis pipeline, and use webhooks to continue work when a task finishes.

Whisper API pricing

$0.01 / minute

Transcription and audio-to-English translation use the same rate. Duration is measured by the second and rounded up to the next whole audio second. Add credits with a one-time purchase.

Website Unlimited is a separate subscription and is not included in API credit packs. There are no automatic top-ups or overage charges. Applicable taxes appear at checkout.

How API credit billing worksWebsite subscription pricing

Private API records and usage controls

Developer API uses cloud processing. Your API records and credits are managed separately from recordings and subscriptions in the website dashboard.

Read the API data guide

  • Control access

    Create scoped API keys with expiry dates and revoke them from your Developer Console.

  • Control each request

    Use an idempotency key for retries and an optional credit limit to bound the cost of a transcription task.

  • Control your results

    Download results before their stated expiry, or delete a completed or failed task's result through the API.

Whisper API FAQ

What is the Whisper API from Whisper Web?

Whisper Web provides a managed Whisper speech-to-text API for audio and video files. Submit a recording, track the asynchronous task, and retrieve transcript text, segment timestamps or subtitle files. You manage requests and API credits through the Developer Console.

Which languages does the transcription API support?

The API supports 99 source languages. Use language: auto for automatic detection or a supported code such as en, es, fr, de, ja or zh. The API reference lists the complete set of accepted language codes.

What are the audio and video file limits?

Each file can be up to 5 GB (5,000,000,000 bytes) and 10 hours long. The server validates the media before transcription. Each Developer account allows 20 nonterminal tasks, with up to 2 concurrent validation executions and 2 concurrent transcription executions.

Which transcript and subtitle formats can I download?

Choose JSON, TXT, SRT or VTT. JSON includes transcript text and segments with start_ms and end_ms timestamps. TXT provides plain text; SRT and VTT provide timed subtitle cues.

Can the API translate audio into English?

Yes. Set task to translate_to_english to produce English text from the source audio. Set task to transcribe to retain the spoken language. Both tasks cost $0.01 per minute, billed by the second.

How much does the Whisper API cost?

Transcription and audio-to-English translation cost $0.01 per minute. Usage is rounded up to the next whole audio second. Buy prepaid packs of $10 for 1,000 minutes, $25 for 2,500 minutes, $100 for 10,000 minutes. There are no automatic top-ups or overage charges. Applicable taxes are shown at checkout.

How do asynchronous transcription and webhooks work?

A new transcription request returns HTTP 202 and a task ID. Query that task until its status is completed or failed, or register a webhook to receive a signed event. After completion, fetch the result through the API. A task ID confirms acceptance, not that processing has finished.

How are API audio and transcripts stored?

The API uses cloud processing. Keys can access only their own Developer account, and API records are separate from website recordings. Temporary source media is cleaned up after a confirmed terminal state. Results expire 30 days after task creation; download them before expires_at or delete a completed or failed task earlier.

Do I need a website Unlimited subscription to use the API?

No. Open a Developer account with your Whisper Web login and purchase API credits separately. Website Unlimited is a separate subscription and is not included in API credit packs. Website subscriptions do not provide API credits.

Is Whisper Web the same service as OpenAI's Whisper API?

Whisper Web is an independently operated service that uses Whisper-based speech recognition. This API uses Whisper Web endpoints, API keys and billing. An OpenAI API key is not used to authenticate Whisper Web API requests.

API documentation and developer resources

Connect an AI agent

Add the Whisper Web MCP server to Claude, ChatGPT, Claude Code, Codex or Cursor and sign in once.

Transcribe directly in your browser

Use Whisper Web's website tools when you want to work with a recording without writing an integration.