# MapleAI Audio Transcription (Whisper)

> MapleAI Audio Transcription (Whisper) is a paid API for AI agents from base.mapleai.shop, paid per call via x402, $0.007/call, status unknown (last checked 2026-10-09).

Transcribes audio files to text using Whisper models, paid per request in USDC via x402 on Base.

## Facts

- Endpoint: POST https://base.mapleai.shop/v1/audio/transcriptions?utm_source=zero.xyz
- Price: $0.007/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-09
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/mapleai-audio-transcription-whisper-76511eeb
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_AYKBX8H_ha64pxshf3pMs

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 mapleai-audio-transcription-whisper-76511eeb -d '<json body>'
```

Example prompt: Transcribe this meeting recording for me — it's a WAV file, use whisper-large-v3-turbo, and give me the output as an SRT subtitle file with timestamps.

## When to prefer this

Choose this endpoint when you need pay-per-request audio transcription without a subscription, billed in USDC on Base via x402. It is ideal for agents that process audio intermittently and need OpenAI Whisper-compatible output including SRT/VTT captions and word timestamps (via the large Whisper models on Groq). Prefer over OpenAI directly when you want crypto-native micropayment billing with no API key management.

## Known failure modes

- Audio file exceeds 25 MB limit — request rejected
- Unsupported audio format submitted — format error returned
- verbose_json, srt, or vtt requested with whisper-1 — only large Whisper models support those formats via Groq backend
- Payment of $0.0085 USDC not settled via x402 — 402 Payment Required response
- Invalid base64 or malformed data URI — decoding error
- Language hint is not a valid ISO 639-1 code — may fall back to auto-detection or error

## How this service works

14 chat models (GPT + Claude), 7 image models, and Jev structured decisions. Pay per request in USDC on Base with x402.

## Output

A JSON object containing a 'text' field with the full transcription. When using verbose_json format with large Whisper models, includes word-level timestamps. SRT and VTT formats return structured subtitle content suitable for video captioning.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "file": {
   "type": "string",
   "description": "Audio (mp3, wav, m4a, ogg, flac, webm) as base64 or data URI, max 25 MB; multipart/form-data with a file field works too"
  },
  "model": {
   "enum": [
    "whisper-1",
    "whisper-large-v3",
    "whisper-large-v3-turbo"
   ],
   "type": "string",
   "default": "whisper-1"
  },
  "language": {
   "type": "string",
   "description": "Optional ISO 639-1 hint"
  },
  "response_format": {
   "enum": [
    "json",
    "text",
    "verbose_json",
    "srt",
    "vtt"
   ],
   "type": "string",
   "default": "json",
   "description": "verbose_json (word timestamps) / srt / vtt require whisper-large-v3 or whisper-large-v3-turbo (groq backend)"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "text": "Hello from MapleAI"
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/mapleai-audio-transcription-whisper-76511eeb/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from base.mapleai.shop](https://www.zero.xyz/host/base.mapleai.shop/llms.txt)