# JarvisClaw Video Transcription API

> JarvisClaw Video Transcription API is a paid API for AI agents from api.jarvisclaw.ai, paid per call via x402, $0.115/call, status unknown (last checked 2026-10-12).

Transcribes any video URL to text using Whisper v3 large, with auto-detection of 90+ languages, optional English translation, and speaker diarization.

## Facts

- Endpoint: GET https://api.jarvisclaw.ai/v1/marketplace/api/video-transcribe?utm_source=zero.xyz
- Price: $0.115/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-12
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/jarvisclaw-video-transcription-api-20191a24
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_xybbAn-dByze5MOnCayti

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 jarvisclaw-video-transcription-api-20191a24
```

Example prompt: Can you transcribe this video for me — https://example.com/interview.mp4 — and identify the different speakers? Also translate everything to English if it's not already in English.

## When to prefer this

Choose this endpoint when you need to convert spoken audio from a video URL into text, especially when the language is unknown or non-English, when speaker identification matters, or when English translation of foreign-language video content is required. It is ideal for videos up to 60 minutes. Prefer this over generic ASR APIs when you need Whisper v3 large quality with built-in diarization and multilingual support in a single pay-per-call call via USDC.

## Known failure modes

- Video URL is unreachable or returns a non-200 response — transcription fails with an access error
- Video exceeds 60-minute or 500MB limit — request rejected with size/duration error
- Unsupported audio codec or corrupt video file — transcription returns an error or empty result
- Language not in the 90+ supported set — detection may fall back incorrectly or fail
- Network timeout fetching large video files — partial or failed transcription
- x402 payment not fulfilled — request blocked before processing begins

## How this service works

Transcribes any video URL to text with Whisper v3 large. Auto-detects 90+ languages, offers a translate-to-English mode, and optional speaker diarization. 60-min / 500MB max.

## Output

Returns a full text transcript of the video's spoken audio, with auto-detected language, optional English translation, optional per-speaker labels (diarization), and timestamps. Handles videos up to 60 minutes or 500MB.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "query": {
   "type": "string",
   "description": "Request parameters"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/jarvisclaw-video-transcription-api-20191a24/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.jarvisclaw.ai](https://www.zero.xyz/host/api.jarvisclaw.ai/llms.txt)