> ## Documentation Index
> Fetch the complete documentation index at: https://developers.t2000.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Localize an audio clip

> An English clip in, the same clip in another language out. fal Whisper transcribes, DeepL translates, ElevenLabs re-voices it. A full audio round-trip, paid as you go, no API keys.

You have a voice clip in English and want it in Spanish — not subtitles, actual spoken audio. Your agent transcribes the clip, translates the text with context-aware AI, then synthesizes a natural voice reading it back in the target language. \~\$0.14, end to end — audio in, audio out.

<Note>
  This is a **durable gateway demo** — every leg needs a real model: real transcription, real translation, real voice synthesis. A sandboxed client can't fake any of them. The artifact (the re-voiced clip) is re-hosted on our store, so the URL is durable.
</Note>

***

## The prompt

```
Use t2 services. Take this English clip and give me a Spanish version I can play:
https://example.com/clip.mp3
Use fal Whisper to transcribe, DeepL to translate, and ElevenLabs to read it back.
```

***

## What runs

1. `POST /fal/fal-ai/whisper` — transcribe the source clip to text (\~\$0.02)
2. `POST /deepl/v1/translate` — translate the transcript, context-aware (\~\$0.02)
3. `POST /elevenlabs/v1/text-to-speech/:voiceId` — synthesize the translated voice (\~\$0.10)

***

## Run it

### Claude Desktop (MCP)

```bash theme={null}
npm install -g @t2000/cli && t2 init && t2 fund && t2 mcp install
```

Paste the prompt with any audio URL and target language. The agent chains the three calls and hands back the new clip's URL.

### SDK

```typescript theme={null}
import { T2000 } from '@t2000/sdk';

const agent = await T2000.create();
const sourceUrl = 'https://example.com/clip.mp3';
const voiceId = '21m00Tcm4TlvDq8ikWAM'; // any ElevenLabs voice

const transcript = await agent.pay({
  url: 'https://mpp.t2000.ai/fal/fal-ai/whisper',
  method: 'POST',
  body: JSON.stringify({ audio_url: sourceUrl }),
});

const translated = await agent.pay({
  url: 'https://mpp.t2000.ai/deepl/v1/translate',
  method: 'POST',
  body: JSON.stringify({
    text: (transcript.body as { text: string }).text,
    target_lang: 'ES',
  }),
});

const speech = await agent.pay({
  url: `https://mpp.t2000.ai/elevenlabs/v1/text-to-speech/${voiceId}`,
  method: 'POST',
  body: JSON.stringify({
    text: (translated.body as { translations: { text: string }[] }).translations[0].text,
    model_id: 'eleven_multilingual_v2',
  }),
});

// The gateway re-hosts the audio on its artifact store — speech.body.url is durable.
console.log((speech.body as { url: string }).url);
```

***

## Expected output

```
3 calls · ~$0.14 · ~12s · 0 taps
A Spanish-voiced .mp3 of the original clip, on a durable URL
```

***

## Extend it

* Target any of DeepL's 30+ languages by changing `target_lang` (`FR`, `DE`, `JA`, …)
* Swap **DeepL** for **Google Translate** (`/translate/v1/translate`) for 130+ languages
* Generate captions too: feed the transcript to **Claude** (`/anthropic/v1/messages`) for an SRT
* Loop over several `target_lang` values in parallel to ship one clip in five languages