▶TranscriptReel
Home / Tools / Video to Audio

Video to audio converter

Take the audio track out of a video file and save it on its own. Drop an MP4, MOV or WEBM in, and the browser decodes the audio and writes a WAV you can download. No upload, no account.

No file loaded. MP4 and MOV with AAC audio, and WEBM with Opus, are the ones that almost always decode.

Why extract the audio at all

Speech-to-text only reads the audio track, so the video track is dead weight: a 60-second 720p clip can be a couple of megabytes with a video track and a few hundred kilobytes as compressed audio. Sending the whole video also means the tool has to demux it first, which is where most "unsupported file" errors come from when you point a transcription service at a screen recording.

Extracting first has a second use: you can listen to just the audio while reading a transcript, and you can trim the silent intro that a screen recorder adds before anything is said. That silence is a common cause of a transcript that comes back with stray words or nothing at all — see what to do when a transcript comes back empty.

The output here is uncompressed 16-bit PCM WAV at the source sample rate. That is on purpose: a second lossy encode before transcription costs accuracy for no benefit. If you need a small file for sharing, convert the WAV afterwards.

Why does a 1 MB video become a 10 MB WAV?

Because WAV stores every sample. A minute of 44.1 kHz stereo audio at 16 bits is 60 × 44,100 × 2 channels × 2 bytes = 10,584,000 bytes plus a 78-byte header. Compressed audio in the same video is maybe a tenth of that. Choose mono output to halve the WAV size without hurting speech recognition.

Can I extract from a file with two audio tracks?

Not reliably. The browser decodes the track it considers default. A dual-language MKV often comes back as whichever track the container marks first. If you need a specific track, demux with a desktop tool such as ffmpeg first: ffmpeg -i in.mkv -map 0:a:1 out.m4a.

Is the video re-encoded?

No. The video track is never read into the output. Nothing is re-encoded at all — the samples you get are the samples that were in the file, only unpacked into WAV.

Is my video uploaded to a server?

No. The file is decoded inside your browser tab with the Web Audio API. Nothing is sent to a server and nothing is stored.