50% off yearly plansLimited time — ends in 50% off yearly plans · ends in

About this skill

A lyric video lives or dies on timing: a line that comes in late, or a word gone before you can read it, is all a viewer notices. This skill measures everything first. It splits the vocal from the song with Demucs, finds the beat, bars and sections with librosa, and times every word of your lyric sheet to the voice with WhisperX's aligner, then moves each word that follows a pause onto the moment the singer comes in. It writes one timing file, and the scenes and type styles read only that. Four scenes (drifting dust, falling light, beat rings and a three.js tunnel) and three type styles (karaoke fill, stacking words and one big word at a time) are mapped to sections in a plan you edit. A renderer draws every frame in headless Chrome and muxes the song, and a checker fails the video when a line appears off its sung time or a word is on screen too briefly to read.

You say

Using this skill, make a vertical lyric video for my song track.wav with the words in lyrics.txt. Calm drifting scene in the verses, the tunnel in the chorus, one big word at a time in the chorus, and check the sync before you show me.

How it works

  1. Give your agent the song, the lyric sheet and the look you want. It checks the song is yours to use.
  2. It splits off the vocal, measures the beat, bars and sections, and times every lyric word to the voice.
  3. It shows you the sections and a timing table, maps each section to a scene and a type style, and renders a still of each.
  4. It renders the video with the song under it, runs the checker, and fixes what fails before you see it.

Getting a good result

  • Write the lyric sheet as it is sung, ad-libs and repeats included, one sung line per line.
  • Use a vocal stem from the mix if you have one. It aligns better than any separated one.
  • Keep scenes calm where the words go, and save the brightest scene for the chorus.

What you get

  • Scripts that split the vocal, measure beats, bars and sections, and time each lyric word to the voice
  • One timing file that the scenes and type styles read, so a new look can't break the sync
  • Four scene templates, three of them Canvas and one three.js, and three kinetic type styles
  • A renderer that draws exact frames in headless Chrome and muxes the song into an MP4
  • A checker that fails lines off their sung time, words too brief to read and a shifted audio track
  • A short song of our own, with lyrics, to try it on

Good fit

  • A lyric video or visualizer for a song you wrote, commissioned or generated
  • Vertical lyric clips of a chorus for short-form posts
  • Karaoke-style versions with each word lit as it is sung

Not for

  • Songs you don't have the rights to, or lyrics you didn't write
  • Videos that need filmed footage or generated video clips
  • Heavily processed vocals the aligner can't follow, which need hand timing

How to install it

  1. Download the ZIP and unzip it.
  2. Move the folder holding SKILL.md into your agent's skills folder, such as ~/.claude/skills/ for Claude Code.
  3. Start a new session and ask for what you want to make.

Needs: An agent that reads skills, Python 3.12 with uv, and about 1.5 GB for packages and models, Node.js 22 or later, with Playwright and Chrome or Chromium, ffmpeg.

Popular on X

Posts showing demand for this kind of video. Unless a post names this tool, it wasn't made with it.

FAQ

What does it cost?

$2.99 once for this skill, or $99 once for every skill in the library. It also comes with every paid AdAnt plan, from $15 a month. It runs on your own computer. Only the optional demo voice uses AdAnt credits, under one, and only if you make it with AdAnt's voice tool.

Can I use it in commercial work?

Yes. The skill's code is MIT licensed: use it and what it makes in your own commercial work, including for clients, and keep the license file with any copy you pass on.

Does it write or look up the lyrics?

No. You give it the lyric sheet, and it never transcribes, so no word reaches the screen that isn't on your sheet. It times your words to the singing. Use it for songs you have the rights to.

How exact is the timing?

On our demo song, the aligner alone put four of six lines 70 to 110 ms late. Moving each line onto the moment the voice comes in brought all of them within 10 ms, and within 30 ms on a spoken-word test with lines off the beat. Both were short, clean vocals of our own; it is untested on a busy commercial mix. The checker then measures the finished video: every line must appear no later than one frame after its first word is sung.

How is this different from Kinetic Type Film and WhisperX Captions?

Kinetic Type Film writes the words first and makes the music fit a beat grid. Here the song comes first and the words follow the singer. WhisperX Captions times speech; sung words need the extra steps this skill adds, and Librosa Beat Sync's approach finds the beat underneath.

Can I make it vertical, or change the look?

Yes. Set the size in the plan, pick a scene and a type style for each section, and set your colours and font. Scenes and type styles are small JavaScript files you can copy and change, and they never decide timing, so the checks still hold.

Did the people who posted these videos write this skill?

No. AdAnt saw lyric videos made in code on X, then wrote this skill, its scripts and its demo song from scratch. Nothing from the posts is copied, and their authors did not write or review it. The clip on this page was made by AdAnt with this skill.

Related skills

$2.99

librosa Beat Sync

Measure a track's beats with librosa, drive an animation from the JSON it writes, and prove the sync by measuring the video.

$2.99

Narrated Explainer Video

Have your agent research a topic, write a script with a source for every figure, voice it, and animate scenes that land on the words that name them.

$2.99

Kinetic Type Film

Make a launch film of big words cut to the beat, with real motion blur, synthesized music and measured sync.

$2.99

Logo Reveal and Ident

Have your agent turn your SVG logo into a 3 to 8 second reveal with a sound hit, as wide, vertical, square and transparent videos that end on your logo, checked against the SVG.

$2.99

WhisperX Word Captions

Get a timestamp for every spoken word on a plain CPU, export SRT and VTT, and drive word-highlighted captions from the JSON.

$2.99

Visible Music Machine

Write a tune as notes first, then have your agent build a machine that plays it, where every note is a ball or mallet hit you can see, with sound on the exact frame.

$2.99

Dreamlike Voyage Browser Game

Have your agent build a calm ride through several invented worlds in three.js that plays in a browser tab and on a phone, check it, and record a clip of it.

Free

JavaScript Animation: Every Frame Drawn in Code

Have your agent draw a short film on an HTML canvas, write its music in code, and render an MP4 with no assets.

$2.99

Code-Rendered 3D Animated Short

Have your agent write a tiny wordless story, build its characters and set in three.js, shoot it from a shot list, score it and render a frame-exact MP4.

Free

Remotion Video

Make video as React components, where every frame comes from the frame number, and render one template for any data.

$2.99

Procedural Character Performance

Animate an original character dancing, walking or trotting to a beat entirely in code, with feet that never slide and every hit on the beat.

$2.99

Parallax Plates Site

Have your agent split one picture into depth plates and build a page that flies through them on scroll, then prove in Chrome that no plate edge ever shows.