Overview
Turns up to 30 s of recorded conversation into a verbatim protocol: time-stamped lines per speaker (one speaker per audio channel), interruptions, pauses, and lines marked unclear or inaudible, plus a ready-to-read protocol text. Speakers come from audio channels: with a two-channel recording (e.g. a call or an interview with one microphone per person) each channel is one speaker; identical channels or a mono file give one speaker (no voice-based diarisation).
Protocol
x402 (POST)
Price
$0.02
Payment Asset
USD Coin
Payment Network
eip155:8453
Category
uncategorized
Provider
ton
Indexed
2026-11-21 08:10:21
Health
Status
unknown
Latency (p50)
—
Uptime (30d)
—
Reliability Score
—
Last Checked
Never
Last Healthy
Never
Consecutive Failures
0
Input Schema
{
"body": {
"audio_base64": "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",
"speaker_names": [
"Host"
]
},
"bodyType": "json",
"method": "POST",
"type": "http"
}
Output Schema
{
"example": {
"limitations": [
"Speakers are separated only by channel; voices mixed into one channel are reported as one speaker.",
"Whisper base can mishear names and numbers and may skip filler words (uh, um), so the protocol is a draft that needs human review, not a certified transcript.",
"At most 30 s and two channels per call."
],
"result": {
"audio": {
"bytes": 1197,
"channels": 1,
"codec": "opus",
"duration_s": 1.42,
"sample_rate": 48000,
"sha256": "3af127aff1c1cf5d827731c34d4476c1d820bde2428840e9500b303c5905cd63"
},
"discarded_segments": 0,
"interruptions": [],
"languages": [
{
"language": "en",
"speaker": "Host"
}
],
"lines": [
{
"avg_logprob": -0.729,
"channel": 1,
"end": 1.3,
"inaudible": false,
"speaker": "Host",
"start": 0.35,
"text": "Hello, hello.",
"unclear": false
}
],
"model": {
"files": [
"onnx/encoder_model_quantized.onnx",
"onnx/decoder_model_merged_quantized.onnx"
],
"license": "MIT (OpenAI Whisper weights and code); ONNX export Xenova/whisper-base under Apache-2.0 per its model card",
"model": "OpenAI Whisper base (74 M parameters)",
"onnx": "https://huggingface.co/Xenova/whisper-base/tree/64da57285918e20ea79ea5c88eed7197933abaa8",
"runtime": "onnxruntime-node 1.31 (CPU), greedy decoding",
"vad": "Silero VAD v6.2"
},
"notes": [],
"pauses": [],
"protocol_text": "[00:00.3] Host: Hello, hello.",
"speakers": [
{
"channel": 1,
"lines": 1,
"name": "Host",
"speech_s": 1.07
}
]
},
"service": "verbatim-protocol",
"status": "success"
},
"type": "json"
}