How can we help?

An LRC file is a file format used to store synchronized song lyrics. Each line begins with a timestamp in [mm:ss.xx] format, followed by the lyric text for that timestamp.

We use our AI transcription engine on your audio file. The AI identifies what is being sung and when, timing the result to the word rather than just the line — so you can drive karaoke-style word-by-word highlighting, not only line changes.

We won't quote a single headline percentage, because the honest position is more specific. We benchmarked our self-hosted fallback transcriber against Jam-ALT (the peer-reviewed ISMIR 2024 sung-lyrics benchmark) and measured a 60% word error rate at its default settings across 20 songs in English, German, Spanish and French, on CPU. That is our worst configuration and we're publishing it rather than our best one (49%, which needs you to tell it the language). Note word error rate is not 100% minus accuracy: that same run got 63% of the sung words right, and the error rate is inflated mainly by the model adding words nobody sang. But most jobs don't use that engine at all: they run through a cloud model we haven't benchmarked yet, so quoting any of these as your expected result would be unsourced. We'd rather say that than pick a flattering number. Context worth having: on the same benchmark a human transcriber scores 11.1% word error and the best commercial system 16.1% — sung lyrics are far harder than speech. Practical tip that we measured: set the Language dropdown instead of leaving it on Auto. Auto guessed wrong on 5 of our 20 test songs, and on those songs the transcription roughly falls apart. Results vary a lot with your track, every generated LRC is fully editable, and fixing a few words beats syncing from scratch.

Only as long as they're useful to you. The instant tools (format conversion, tag and artwork editing, noise cleanup) delete your file the moment the result is returned. AI jobs keep the original a little longer so you can replay it from Job History, then delete it automatically. For client-side tools (converters, validators, viewers), files never leave your browser.

We support MP3, WAV, FLAC, AAC, OGG, and M4A. For best results, use MP3 or WAV at 44.1kHz.

Standard LRC has line-level timestamps (one per line). Enhanced LRC has word-level timestamps, enabling karaoke-style word-by-word highlighting.

Free and Starter plans are for personal use. Premium and Elite plans include commercial licensing.

The AI LRC Generator and Lyrics Extractor support 57 languages via our AI engine, including English, Spanish, French, German, Japanese, Korean, Chinese, Arabic, and more.