What does it cost to convert MP3 to WAV?

$0.003 per call, paid in USDC on Base via x402.

How large can the MP3 file be?

Up to 100 MB (104,857,600 bytes), and at most 60 min (3,600 s) of media. Past a limit the call answers too_large (413).

What happens if the conversion fails?

The tool answers with a typed JSON error from the errors table. Payment is settled only after the tool has produced its result; a call that fails inside the tool is never settled.

catalogue / tools / Audio & video / MP3 to WAV

Convert MP3 to WAV

To convert MP3 to WAV, send one POST to /v1/audio-convert with the MP3 source and "to": "wav" in the body. Left out, "to" falls back to its default "mp3", so a WAV request has to send it. Each call costs $0.003 in USDC on Base via x402, with no account and no API key. The source file can be up to 100 MB and 60 min (3,600 s) long.

Endpoint: POST /v1/audio-convert
Price: $0.003 per call
Service: audio-convert
Docs: /docs/audio-convert

01Limits

Limit	Value
Largest source file	100 MB (104,857,600 bytes)
Longest input	60 min (3,600 s)
Timeout	94 s
Source formats	AIFF, AVI, FLAC, MKV, MP3, MP4, Ogg, WAV
Output formats	AAC, FLAC, MP3, Opus, WAV

02Example

A real MP3 to WAV run of audio-convert.

Request

POST /v1/audio-convert
content-type: application/json
payment-signature: <base64 x402 payload>

{
  "url": "https://grist.tools/samples/av/pair-tone.mp3",
  "to": "wav"
}

Response

200 OK
payment-response: <base64 settlement receipt>

{
  "source_url": "https://grist.tools/samples/av/pair-tone.mp3",
  "bytes": 12980,
  "source_container": "mp3",
  "format": "wav",
  "bitrate_kbps": null,
  "sample_rate": null,
  "audio_bytes": 48044,
  "audio_base64": "UklGRqS7AABXQVZFZm10IBAAAAABAAEAQB8AAIA+AAACABAAZGF0YYC7AABg40Po..."
}

Source file: av/pair-tone.mp3, 12.7 KB (12,980 bytes). 3 s mono 8 kHz MP3 at 32 kbit/s, encoded by ffmpeg's libmp3lame with the bit-exact flags: three 0.5 s 500 Hz triangle beeps at rising level, each followed by 0.5 s of silence. The recorded run returned 46.9 KB (48,044 bytes) of WAV, 3.7 times the source size, with the source read as the mp3 container. Decoded, the WAV in audio_base64 opens with 52 49 46 46 A4 BB 00 00 57 41 56 45 (ASCII RIFF and WAVE). Recorded with ffmpeg 8.1.2.

In the sample itself: the file opens with an ID3v2.4 tag with 10 bytes of body, then an MPEG 2.5 Layer III frame header at byte 20: 32 kbps, 8,000 Hz, mono. That first frame is an Info header rather than audio, and it counts 44 audio frames and 12,960 bytes. In the returned WAV: the RIFF header counts 48,036 bytes after itself, and the fmt chunk, in its own field order, declares audio format 1 (integer PCM), 1 channel, 8,000 Hz, 16,000 bytes per second, a block align of 2 bytes and 16 bits per sample. The data chunk right after it holds 48,000 bytes of samples, 3 s at that rate.

03Format notes

MP3 as a source

MP3 is lossy audio. Grist accepts it when the file opens with an ID3v2 tag or with a bare MPEG audio frame sync: the byte FF followed by a byte whose top three bits are set. That sync test also matches a raw ADTS AAC stream, and because nothing proves ffmpeg would then pick the right demuxer, AAC is kept out of the sources. Duration needs care with MP3. For a file without a Xing header, the probe estimates length by dividing file size by the bitrate of the first frames, which a crafted variable-bitrate file can falsify, so any probe warning sends the measurement to a scan of real packet timestamps. Once accepted, the first audio stream is decoded by ffmpeg; a lossless destination stores the decoded samples without recovering what MP3 encoding discarded.

WAV as a target

WAV output is written with the pinned pcm_s16le codec, so every result holds signed little-endian PCM regardless of the source encoding. The WAV muxer fills in its RIFF chunk sizes when the file closes, so Grist points ffmpeg at a seekable temporary file inside the per-call directory and reads it back, rather than streaming to a pipe. A bitrate_kbps value is rejected for this lossless destination instead of being ignored. The optional sample_rate must come from the service's closed set; left out, ffmpeg chooses, normally keeping the source rate. Only the first audio stream is mapped, with video, metadata and chapters removed, and bit-exact flags make the bytes reproducible for the pinned ffmpeg. PCM from a lossy source contains the decoded samples at full size, without restoring detail the earlier codec discarded.

04Errors

Code	HTTP
`invalid_input`	`400`
`blocked_target`	`403`
`unreachable_target`	`424`
`upstream_timeout`	`424`
`unsupported_content_type`	`415`
`too_large`	`413`
`unprocessable`	`422`
`internal`	`500`

When each is raised, and what it means for payment: /docs/audio-convert.

05Questions

What does it cost to convert MP3 to WAV?: $0.003 per call, paid in USDC on Base via x402.
How large can the MP3 file be?: Up to 100 MB (104,857,600 bytes), and at most 60 min (3,600 s) of media. Past a limit the call answers too_large (413).
What happens if the conversion fails?: The tool answers with a typed JSON error from the errors table. Payment is settled only after the tool has produced its result; a call that fails inside the tool is never settled.