# Fish Audio Transcriptions via OpenRouter

> Fish Audio Transcriptions via OpenRouter is a paid API for AI agents from Fish Audio at fish-audio.mpp.tempo.xyz, paid per call via MPP, price unknown, status unknown (last checked 2026-10-27).

Transcribes base64-encoded PCM WAV audio (mono, 16-bit, 16 kHz, up to 10 minutes) into text using Fish Audio speech-to-text models.

## Facts

- Endpoint: POST https://fish-audio.mpp.tempo.xyz/v1/audio/transcriptions?utm_source=zero.xyz
- Price: price unknown
- Payment: MPP
- Status: unknown
- Last checked: 2026-10-27
- Activations on Zero: 0
- Provider: Fish Audio
- Docs: https://openrouter.ai
- Website: https://openrouter.ai
- Tags: ai
- Canonical page: https://www.zero.xyz/c/fish-audio-transcriptions-via-openrouter-5fa3d6c3
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_OpLtEfc5p2vIaSkI1dzwZ

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability fish-audio-transcriptions-via-openrouter-5fa3d6c3 -d '<json body>'
```

Example prompt: Can you transcribe this WAV audio recording for me? It's a mono, 16-bit, 16 kHz file I'll send as base64 — it's about 3 minutes long and I need the text transcript back.

## When to prefer this

Choose this endpoint when you need fast, on-demand speech-to-text transcription of short-to-medium audio clips (up to 10 minutes) in PCM WAV format via a one-shot micropayment model. It is especially useful for AI agents that need to transcribe audio programmatically without a subscription, leveraging Fish Audio's models through the OpenRouter payment infrastructure.

## Known failure modes

- Audio not in expected PCM WAV format (mono, 16-bit, 16 kHz) causes rejection or poor results
- Audio duration exceeding 10 minutes may be rejected or truncated
- Invalid or malformed base64 encoding causes a parsing error
- Unsupported or incorrect model ID results in a model-not-found error
- Background noise or low audio quality may degrade transcription accuracy
- Payment or billing failure if USDC balance is insufficient

## How this service works

Transcribe base64 PCM WAV (mono, 16-bit, 16 kHz, up to 10 minutes; models and prices at GET /v1/audio/models)

## Output

Returns the transcribed text content extracted from the submitted audio, representing the spoken words detected in the WAV audio input.

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/fish-audio-transcriptions-via-openrouter-5fa3d6c3/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from fish-audio.mpp.tempo.xyz](https://www.zero.xyz/host/fish-audio.mpp.tempo.xyz/llms.txt)