Decoding uses your browser's own audio decoders, so any file your browser can play will convert here. Exact support varies by browser.
MP3
WAV
FLAC
M4A / AAC
OGG / OGA
OPUS
WEBM / WEBA
AIFF / AIF
MP4
M4V
MOV
Video files are read for their audio track only; the picture is ignored. MKV is not supported — browsers cannot decode it natively.
How it works
Your file is decoded in the browser and turned into a polyphonic MIDI track using Basic Pitch, a small neural model from Spotify's Audio Intelligence Lab. It detects several notes sounding at once, keeps their start times and lengths, estimates the tempo, and writes a standard .mid file you can open in any DAW or editor (Ableton, FL Studio, Logic, GarageBand, MuseScore, and others). The model is about 1 MB and is downloaded once, then it runs on your device. Nothing is uploaded — the audio never leaves your browser.
An audio to MIDI step like this is transcription, not format conversion: the output is a list of notes, not a recording. A WAV to MIDI or FLAC to MIDI job goes through the same path; uncompressed files tend to give the cleanest note detection.
Known limitations
The transcription is polyphonic (several notes at once), but it does not separate instruments into different tracks. Vocals, bass, drums and other parts land in one track; splitting them per instrument is source separation, which this page does not do.
Basic Pitch works best on one instrument at a time. On dense, fully mixed songs it will still find notes but can miss or misplace some.
Percussion and noise do not produce meaningful pitched notes and are mostly ignored.
Tempo is estimated from the note onsets; when no steady pulse is found the file falls back to 120 BPM and the stats panel says so.
The first conversion downloads the ~1 MB model, so it is slower than later runs. Processing time scales with file length.
Length limit: up to 10 minutes or 200 MB per file. Longer audio is refused with a message instead of being processed, because a full-length file would exhaust browser memory. Audio is transcribed in 30-second segments inside a background worker, so peak memory stays flat and the page stays responsive.
Processing runs slower than playback — roughly 3× the audio length in our own tests (a 30-second clip took 96 seconds, a 2-minute 22-second clip took 395 seconds). The percentage and a time estimate are shown while it runs.
Frequently asked questions
Is this MP3 to MIDI converter free?
Yes. There is no account, no email and no credit card. The transcription runs on your own device, so there is no server bill to pass on.
Where does my file go?
Nowhere. The audio is decoded and analyzed inside your browser tab. It is never sent to a server.
Can it convert WAV, FLAC or OGG to MIDI too?
Yes — this is also a WAV to MIDI, FLAC to MIDI and OGG to MIDI converter in the same box, along with M4A, OPUS, WEBM and AIFF. If your browser can play it, this page can read it.
Does it work with video like MP4 or MOV?
Yes for the audio track. Drop an MP4, M4V or MOV file and the page reads its sound and converts that. MKV is not supported because browsers cannot decode it.
How accurate is the conversion?
It depends on the source, and this page does not invent a number for you. Instead it shows the result before you download: the note count, the pitch range, the tempo and a piano-roll preview. If the notes look wrong for your file, you know immediately rather than finding out later in your editor.
What tempo is written into the MIDI file?
The tempo is measured from the note onsets and written into the file; note positions are computed from that same tempo so they land at their true times. When no steady pulse can be found the file falls back to 120 BPM and the panel tells you that happened.
Is there a file length limit?
Yes — 10 minutes or 200 MB, whichever comes first. Longer files are rejected with a message rather than left to crash the tab. The reason is memory: the transcription model holds intermediate note data in the browser, and a full album side is past what a tab can safely keep. If your file is longer, cut the section you need and convert that.
How long does a conversion take?
Roughly 3× the audio length in our test environment: 30 seconds of audio took 96 seconds, and a 2-minute 22-second file took 395 seconds (6.6 minutes). A machine with GPU acceleration is faster; a phone is slower. Because the work runs in a background worker, the page stays usable and the status line shows a percentage with a time estimate.
I cannot read sheet music — how do I check the result?
Press Play result. It plays the transcribed notes back through your browser with a plain tone, so you can compare it by ear with the original: does the melody go the same way, does the rhythm line up, is anything missing or added. If it sounds right, download the .mid and open it in an editor.
Does it work on a phone?
Yes. The page is built mobile-first — the drop zone, buttons and note list all fit a phone screen. Converting a long file on a phone is slower than on a laptop, which is expected.