Lyrics Transcription from Audio — song words from a recording

An online tool that recognises song lyrics from an audio recording. It separates the vocals from the music, finds the parts where someone sings, picks a model for the language and returns the words line by line with timestamps — for karaoke, subtitles, study and translation.

How to get the lyrics of a song

1

Upload the song — audio or video up to 15 minutes.

2

Pick the language or keep auto-detection; switch the model to “More accurate” if you like.

3

Press “Recognise the lyrics” and wait for the stages to finish — progress is shown per stage.

4

Check the lines: doubtful spots are worth comparing by ear, especially choruses with backing vocals.

5

Copy the text or download TXT, LRC or SRT.

🔥 Trending right now
🗣️ Text to Speech ✨ Video Enhancer ✂️ Trim Audio ✨ Audio Enhancer 🎭 Voice Changer 🎙️ Vocal Remover 🎬 Create Video from Audio 🎬 Trim Video

Recognise the words of any song from a recording

Upload a song to get its lyrics
Up to 15 minutes long
MP4, WEBM, MOV, MP3, WAV, M4A, OGG, FLAC, AAC, M4B, WMA, AIFF, OPUS, CAF, MKV, AVI, WMV, FLV, M4V, 3GP, TS, MTS, M2TS, VOB, MPG, MPEG, OGV ·

“More accurate” uses a bigger model: slower, but noticeably more words come out right. Russian songs also get the vocals separated.

Reading the file…

Processing runs on your device. A 3–4 minute song takes from one to several minutes depending on your computer or phone.

No singing heard in this track

The filter found no vocals — this looks like an instrumental. If there is a voice after all (quiet or heavily processed), you can recognise without the filter; expect extra words.

Song lyrics

    Don't know what to try?
    700+ tools — open a surprise
    🎲 Surprise me

    Features

    Vocals first, music second
    Before recognition the voice is separated from the instruments with the Kim Vocal 2 model. On a full mix speech recognition invents words; on separated vocals it reads them.
    A model per language
    Russian runs on GigaAM, every other language on Whisper. Pick the language by hand or let it be detected from the vocals.
    A filter against invented lines
    Gating on the vocals and a repetition block keep the model from looping over instrumental breaks or inserting phrases the song never had.
    Line-by-line text with timestamps
    Lines follow the sung phrases, each with a start time. Export to TXT, LRC for karaoke players and SRT for video.
    Runs in the browser
    Separation and recognition run through WebAssembly and WebGPU on your own device — nothing to install.

    Frequently asked questions

    Why does ordinary speech recognition not work on songs?

    Speech models expect speech. On a song the speech detector takes singing for music and returns nothing, and large models on a full mix invent phrases like “Thank you for watching”. Here the vocals are separated from the music first, then the sung parts are found, and only those go to recognition.

    How accurate is the result?

    It depends on the language, the style and the mix. On a studio pop recording with clear vocals most lines come out right; rap, choirs, heavy reverb and ad-libs produce more errors. Doubtful lines are worth checking by ear — the timestamp takes you to the spot.

    Which languages are supported?

    English, Russian, Spanish, Portuguese, German, French, Indonesian, Korean, Arabic and the other languages Whisper knows. Russian uses GigaAM — on singing it is more accurate and faster.

    What if the track has no vocals?

    The tool says no singing was found and offers to recognise without the filter — for a quiet or heavily processed voice.

    Can I get a karaoke file?

    Yes: download LRC — timestamped lines are read by karaoke and music players. For video there is SRT.

    💡 Want us to improve this tool just for you?

    We can! Just send us a quick message with your idea. If you'd like to discuss it in detail, leave your email and we'll get back to you. You can stay anonymous.

    You May Also Like

    🎧 Audio Translator 📝 Audio to Text 🔄 Change the voice 🗒️ Dialogue analysis and conversation protocol 🥁 Drum Separator 🎤 Karaoke Maker 📢 Megaphone Effect 🎙️ Online Voice Recorder

    How do you rate this tool?

    Thank you for your rating!
    Want to share more? Leave a comment!
    Thank you! Your comment will appear after moderation.
    Who is this tool for?
    🎵 Musicians & Producers 🎬 Video Creators ✨ Content Creators 🎓 Students ♿ Accessibility Users
    Published Updated Author: