LyricVideoMaker

ON AIR · FREE · NO SIGN-UP

AI lyric video generator

AI listens to the vocals, writes down each word and when it is sung, and marks the few it is unsure of.

Or from Suno

Your song is processed to time the words and is not stored or used to train AI.

Public Suno songs only. Use songs you made, and check Suno’s terms before using a video commercially.

How sung words become timed text

Speech models are built around talking, and singing is a harder input. Vowels get stretched across a whole bar, consonants disappear under the snare, and a long reverb tail can sound like an extra syllable. So nothing is transcribed from the full mix. The vocals are separated from the music first, and the AI listens to a voice track with the drums, bass and pads taken out.

That isolated vocal goes to WhisperX, which writes down each word along with its start and end time. Word-level timing is what makes the Karaoke fill and the one-word-at-a-time Kinetic style possible. With only line-level timing you'd know when a line begins and nothing more.

Pasting your lyrics changes the job. The AI no longer has to guess the spelling of what it hears; it only has to line your words up with the audio. That's why pasted lyrics get timing that is close to perfect.

Where the AI slips, and how you find out

The mistakes cluster in the same spots on most songs: ad-libs and backing vocals that overlap the lead, heavy autotune or distortion, very fast rap verses, notes held for several seconds, and invented words or names. A doubled vocal can also leave a faint second voice in the separated track.

You don't have to watch the whole video to catch these. Words the AI is unsure of are marked, so you go straight to them, click, and fix the text or nudge the timing by a fraction of a second.

Sometimes the lyrics you paste and the take on the record disagree: a chorus sung once more than written, a word changed in the booth. In that case the tool uses what it heard and tells you, so the screen shows what the listener actually hears.

What happens to your audio

The audio is sent once, for transcription. It isn't stored and isn't used to train AI. Everything after that, editing and rendering included, runs in your browser. You get 5 transcriptions per day, so it pays to paste the lyrics before the first run instead of re-running the same song.

● TRACK 01 — STYLES

One song.
Three ways to see it.

Switch the style or the shape at any time. The timing of every word stays put.

we were neon in the rain

running every red light home

Karaoke — lines fill as they are sung16:9 · YOUTUBE

running

every

red light

Kinetic — words pop on the beat9:16 · TIKTOK · REELS

Light is breaking
over the hill

Minimal — calm, line by line1:1 · WORSHIP
● SIDE B — EXAMPLES

Real renders,
not mockups.

Each example is drawn live by the same engine that exports your MP4. Lyrics are original samples.

Suno AI song for YouTube — word-by-word karaoke over a night-city background.Karaoke · 16:9 · YOUTUBEMake one like this →
TikTok / Reels hook — big kinetic words in 9:16, built for the first three seconds.Kinetic · 9:16 · TIKTOK · REELSMake one like this →
Chinese lyrics, character by character — each character lights up on time; lines never break inside a word.Karaoke · 16:9 · BILIBILI · YOUTUBEMake one like this →
Rap verse, stacked — each bar piles onto the last as it lands.Stack · 9:16 · SHORTS · TIKTOKMake one like this →
Worship lyrics — calm, readable lines for a projector or livestream.Minimal · 16:9 · PROJECTOR · LIVESTREAMMake one like this →
Lo-fi loop, square — quiet 1:1 lines for Instagram or a lo-fi stream.Minimal · 1:1 · INSTAGRAM · LO-FIMake one like this →
● TRACK 02 — HOW TO MAKE A LYRIC VIDEO

Three minutes,
start to finish.

  1. 00:00

    Drop the song

    Any MP3, WAV or M4A. Paste the lyrics too if you have them, and the timing gets near perfect.

  2. 00:40

    AI listens

    It writes down every word and when it is sung. Words it is unsure of are marked, so you only check those.

  3. 02:10

    Pick a look, export

    Choose a style and a shape. The video renders right in your browser and downloads as an MP4.

● LINER NOTES — PRICING

Free to export.
Pay per song for HD.

No subscription. Most artists need one video per release, so that is how we charge.

A1 · FREE

$0

  • 720p MP4 download
  • Small watermark in the corner
  • Karaoke and Minimal styles, every shape
  • For personal use
Start free
A2 · ONE HD VIDEO

$9.99

  • 1080p, no watermark
  • OK on monetized channels and ads
  • Unlocked on your account, no email needed
  • All four styles plus stock video backgrounds
B1 · THREE-SONG PACK

$19.99

  • Everything in A2, three times
  • Use on any three songs
  • Credits never expire

Buy HD right from the export screen. Credits stay on your account for any song.

● TRACK 03 — QUESTIONS

Asked
often.

Which AI model times the words?

WhisperX. It runs on the vocal track after the vocals have been separated from the music, which is why it copes with dense mixes better than transcribing the full song would.

The AI wrote a word I never sing. Where did it come from?

Usually a backing vocal, an ad-lib or an effect that survived the vocal separation. It is often one of the marked words. Click it and correct the text, or paste your lyrics so the AI has the real words to work from.

How many songs can the AI transcribe for me?

5 transcriptions per day per person. Edits, style changes and re-exports of a song you've already transcribed don't count against that.

Is this lyric video maker really free?

Yes. Make and download a 720p video with a small watermark in the Karaoke or Minimal style. Pay when you want 1080p without the watermark, the Kinetic and Stack styles, or stock video backgrounds.

How accurate is the word timing?

Most words land within a fraction of a second. Pasting your lyrics makes it near perfect. Any word the AI is unsure of is marked so you can fix it in one click.

Can I use a song I made with Suno or Udio?

Yes. Paste the song's public Suno link and the audio, title and lyrics come in on their own, or download an MP3 or WAV from Suno or Udio and drop it in. You need the rights to the song you use.

What happens to my audio file?

It is sent once to time the words and is not stored or used to train AI. The video itself is made in your browser.

Which shapes can I export?

16:9 for YouTube, 9:16 for TikTok, Reels and Shorts, and 1:1 for feeds. Switch any time; the timing stays.