# CrowdSpark Agent Tools – Audio Transcription (Whisper)

> CrowdSpark Agent Tools – Audio Transcription (Whisper) is a paid API for AI agents from agents.crowdspark.me, paid per call via x402, $0.02/call, status unknown (last checked 2026-11-01).

Transcribes or translates an audio file from a public URL using Whisper large v3 turbo, returning full text, detected language, duration, timestamped segments, and WebVTT captions.

## Facts

- Endpoint: GET https://agents.crowdspark.me/v1/audio/transcribe?utm_source=zero.xyz
- Price: $0.02/call
- Payment: x402
- Status: unknown
- Last checked: 2026-11-01
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/crowdspark-agent-tools-audio-transcription-whisper-c1f5836c
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_z5jruYXWbz0xnXyuEo0Uz

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 crowdspark-agent-tools-audio-transcription-whisper-c1f5836c
```

Example prompt: Can you transcribe this podcast episode for me? Here's the MP3 URL: https://example.com/episode42.mp3 — I'd like the full text plus timestamped segments and captions.

## When to prefer this

Choose this endpoint when you need to transcribe or translate audio from a public URL with rich output including timestamped segments and WebVTT captions, supporting podcasts, voice notes, calls, and meetings in common formats (MP3, WAV, M4A, OGG, FLAC, WebM) up to 10 MB. It is powered by Whisper large v3 turbo and offers automatic language detection plus optional English translation — making it ideal for agents that need structured, segmented transcripts rather than just raw text.

## Known failure modes

- Audio file URL is not publicly accessible or returns 4xx/5xx — transcription fails with an error
- Audio file exceeds 10 MB limit — rejected with a size error
- Unsupported audio format (e.g. AAC in an MP4 container not as M4A) — returns format error
- Very short or silent audio with no speech detected — returns empty text or minimal segment
- Network timeout fetching the remote audio URL — returns a fetch error

## How this service works

Pay-per-call utilities for AI agents over x402, USDC on Base.

## Output

A JSON object containing: the task type (transcribe or translate), full transcript text, detected language code, audio duration in seconds, byte size, model name, audio format, word count, and an array of timestamped segments each with start time, end time, and text. Suitable for display, search indexing, or generating WebVTT captions.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "$schema": "https://json-schema.org/draft/2020-12/schema",
 "required": [
  "input"
 ],
 "properties": {
  "input": {
   "type": "object",
   "required": [
    "type",
    "method"
   ],
   "properties": {
    "type": {
     "type": "string",
     "const": "http"
    },
    "method": {
     "enum": [
      "GET"
     ],
     "type": "string"
    },
    "queryParams": {
     "type": "object",
     "required": [
      "url"
     ],
     "properties": {
      "url": {
       "type": "string",
       "description": "Public URL of the media file itself"
      },
      "vad": {
       "type": "string",
       "description": "true to apply voice activity filtering"
      },
      "task": {
       "enum": [
        "transcribe",
        "translate"
       ],
       "type": "string"
      },
      "language": {
       "type": "string",
       "description": "Optional ISO code, e.g. en"
      }
     }
    }
   },
   "additionalProperties": false
  },
  "output": {
   "type": "object",
   "required": [
    "type"
   ],
   "properties": {
    "type": {
     "type": "string"
    },
    "example": {
     "type": "object"
    }
   }
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "task": "transcribe",
  "text": "Hello and welcome to the show.",
  "bytes": 67000,
  "model": "whisper-large-v3-turbo",
  "format": "mp3",
  "language": "en",
  "segments": [
   {
    "end": 4.2,
    "text": "Hello and welcome to the show.",
    "start": 0
   }
  ],
  "word_count": 6,
  "duration_seconds": 4.2
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/crowdspark-agent-tools-audio-transcription-whisper-c1f5836c/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from agents.crowdspark.me](https://www.zero.xyz/host/agents.crowdspark.me/llms.txt)