music-to-video

Generates beat-synced HyperFrames videos from music tracks using audio analysis and per-frame compositions.

1|Updated Aug 28, 2026
One-click install
npx skills add https://github.com/mooch10/mochiptos --skill music-to-video-mooch10
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: music-to-video
Source: https://github.com/mooch10/mochiptos/tree/main/frontend/.agents/music-to-video
Command: npx skills add https://github.com/mooch10/mochiptos --skill music-to-video-mooch10

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires librosa, numpy, soundfile, and includes scripts (resource) and references (resource) components.

What problem does it solve? Turning a music track into a beat-synced video normally requires manual timeline editing, beat detection, and frame-by-frame motion design. This Skill automates that pipeline: it analyzes the audio once, plans frames against the detected beat grid, and builds each frame as an HTML composition rendered to a final MP4. ## Core Features & Use Cases - Deterministic Audio Analysis: A single analyzer script (librosa-based) produces an audiomap.json with BPM, beat grid, drum classification, energy phases, rolls, silences, and key moments that drives all visual timing. - Orchestrated Frame Pipeline: A six-step workflow (setup, analyze, skeleton, plan, build, assemble, render) cuts the track into frames, dispatches one sub-agent per frame, and assembles a playable index.html. - Template and Motion-Primitive Catalogs: Planners match each frame to reusable templates or free-compose from motion primitives, with optional user images/videos woven in via beat-cut or ken-burns treatments. - Use Case: Give the Skill a synthwave track and a few product photos; it produces a beat-synced promotional video where cuts land on drum hits and the final render is an MP4 matching the track duration exactly. ## Quick Start Use the music-to-video skill to turn my audio file track.mp3 into a beat-synced video with my photos from the assets folder.

Frequently Asked Questions about music-to-video

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I make a beat-synced video from a music track?▼

Provide an audio file (or a video to extract audio from) and the Skill runs analyze-beatgrid.py to detect beats, onsets, and energy phases, then plans frames against that grid and renders an MP4 via the hyperframes CLI. User-supplied images or videos are cut in on the same beat grid.

What audio analysis libraries does the beat detection use?▼

The analyzer uses librosa for beat tracking and onset detection, numpy for numerical processing, and soundfile for audio decoding, with ffmpeg handling format conversion. It classifies drum hits (kick, snare, hihat) via band-split heuristics without machine learning models.

Can I use my own images and videos in the generated video?▼

Yes, user-supplied media is staged into the project's assets folder and woven into frames using beat-cut, ken-burns, or background-under-text treatments. Assets are optional since typography and templates alone can produce a complete video.

Does the beat grid work on calm or ambient music?▼

On calm music the detected BPM grid is treated as unreliable, so frames use phrase_flow pacing driven by phrases and energy envelopes instead of hard beat cuts. The Skill decides this per frame based on rolls, onset density, and energy phases.

What are the limitations of the music-to-video pipeline?▼

The pipeline supports no narration or voiceover in the base version, requires Python audio dependencies and the hyperframes CLI installed, and works best with rhythmic high-energy tracks. Frame count stays small (roughly 1-6) since density is handled inside frames rather than by adding frames.