# Video to MP3 Extractor

> Extract a video’s soundtrack as MP3, WAV or FLAC, or copy the original audio losslessly. Extracts an audio track from MP4, MOV, MKV, WebM, AVI and other videos. Pick the track if there are several (e.g. languages), save it as MP3 (VBR or CBR), WAV or FLAC, or copy the original AAC/Opus/Vorbis/MP3/FLAC stream into M4A/Ogg/MP3/FLAC without re-encoding. Optional start/end trim.

Part of usefulweb.site (https://usefulweb.site/). Results are deterministic calculations from the inputs and the listed assumptions; the API returns exactly what the web page shows.

## Use it
- Human interface: https://videotoaudioextractor.usefulweb.site/ (interactive tool at https://videotoaudioextractor.usefulweb.site/tool)
- JSON API: POST https://videotoaudioextractor.usefulweb.site/api/v1/calculate with a JSON object of inputs (schema: https://videotoaudioextractor.usefulweb.site/api/v1/tool)
- OpenAPI: https://videotoaudioextractor.usefulweb.site/openapi.json
- MCP (Streamable HTTP): https://videotoaudioextractor.usefulweb.site/mcp - tool `video_to_audio_extractor`

## Inputs
- file (file, required): Video file - MP4, MOV, MKV, WebM, AVI, MPEG-TS, WMV, FLV; max 250 MB and 20 minutes. Processed on our server in a private temporary folder that is deleted afterwards; the result can be downloaded for about 2 hours.
- stream (integer): Audio track number - 1 = first audio track. The report lists all tracks with language and codec.
- format (select): Output
- mp3Quality (select): MP3 quality
- start (text): Start time - Optional. Seconds, mm:ss or hh:mm:ss.mmm.
- end (text): End time - Optional. Empty = to the end.

## Method
- ffprobe lists the audio tracks (codec, channels, sample rate, language). The chosen track is mapped with -map 0:a:N; video, subtitles and data are dropped.
- MP3 uses LAME VBR (-q:a 0/2/4) or CBR; WAV is 16-bit PCM; FLAC is lossless level 5.
- "Copy" keeps the original bitstream: AAC/ALAC → M4A, MP3 → MP3, Opus → .opus (Ogg), Vorbis → Ogg, FLAC → FLAC, 16/24-bit PCM → WAV. Other codecs (e.g. AC-3, DTS) need re-encoding and are refused in copy mode.
- Trimming seeks to the start time and stops after the requested length. In copy mode cuts land on audio frame boundaries (a few milliseconds).