Sync Lyrics to Audio

The Generator listens and decides what the words are, then times them. The Aligner is told what the words are and only solves the timing. If you already have correct lyrics, the Aligner is strictly more accurate — there is no transcription step left to make a mistake in.

No. Lines are passed through exactly as you supply them — nothing is deduplicated, filtered, corrected, or dropped. A repeated chorus line stays repeated. The only thing added is a timestamp.

It is the share of your lines the aligner could confidently anchor to the audio. Above 70% is a healthy run. A low score usually means the lyrics don't match this recording (a different version, a cover, or a verse missing) or the track is largely instrumental.

The language picker offers Auto Detect plus English, Hindi, Punjabi, Urdu, Bengali, Spanish, French, German, Chinese, Japanese, Korean, Portuguese and Arabic. For non-Latin scripts your original text is timed as supplied — nothing is transliterated. Naming the language explicitly, rather than leaving Auto, improves timing on non-English songs.

Yes — use Import from file with any LRC, SRT, VTT, ASS or TTML. The existing timestamps are stripped and only the text is kept, then re-timed against your audio. This is the fastest way to fix a lyric file that drifts out of sync.

Use the AI LRC Generator instead — it transcribes and times in one pass. You can always bring its output back here later if you correct the wording.

One credit per job at current pricing — the badge at the top of the page always shows the live price, and the credit is refunded automatically if the job fails. You need to be signed in. Paste up to 20,000 characters of lyrics and upload the song as an audio or video file (MP3, WAV, FLAC, M4A, OGG, MP4, MKV, MOV, WebM and more). Once a job starts you can close the tab — it finishes on our servers and shows up here and in Job History.