Getting Started with the CLI
Install the CLI globally, initialize your workspace, and launch the local Studio in four easy commands.
MotionMind is an open-source skill, CLI and local Studio for making videos with the AI coding agent you already use. The agent writes each video as a HyperFrames project (HTML and GSAP) and renders it to MP4. MotionMind gives the agent a production process, and gives you a record of every version it makes.
This 30-second walkthrough was made with MotionMind itself. It has sound.
The Problem It Solves
AI agents can already write a motion graphic. What goes wrong is everything around it: the agent skips straight to code, every "small change" overwrites the version the client approved, nobody knows which render was the good one, the audio is at a random level, and the next chat starts from zero.
Three Parts
| Part | What it does |
|---|---|
| Skill | Tells the agent how to work: four gates, HyperFrames rules, motion and sound guidance, a QA checklist. |
CLI (motionmind) | Creates versions, renders, measures, logs and promotes them; makes music beds; lists sounds. The agent runs it; you can too. |
Studio (motionmind ui) | A local dashboard at 127.0.0.1:4310: every version with its player, contact sheet, loudness and QA, compare, and Edit in HyperFrames. |
MotionMind doesn't generate video by itself. Your agent does the work, following the skill; that's also why it works with whichever agent and model you prefer. Each version ends up as an H.264 video.mp4, next to its contact sheet and the HyperFrames source that made it.
How a Video Gets Made
You brief the agent. It proposes concepts and waits (Gate 1), writes a storyboard and locks the style in TASTE.md (Gate 2), builds and renders a draft you review in the Studio (Gate 3), and promotes it to production only after QA and your go-ahead (Gate 4).
Every Version on Record
Every render becomes a version: video, contact sheet, source snapshot, loudness, parameters and checksum. Nothing is overwritten. A new idea is a new version, and the one in production is the approved one.
Try It
You need Node.js 22+, ffmpeg on your PATH, and an agent CLI.
npm i -g motionmind
motionmind setup my-studio
cd my-studio
motionmind doctor
motionmind ui
Then, in a second terminal in my-studio, start your agent and brief it:
Use the motionmind skill. I want a 15-second 9:16 promo for my coffee shop "Kopi Senja",
message "great coffee, friendly prices", CTA kopisenja.id, music + SFX, no voice-over.
Where to Go Next
| To learn | Read |
|---|---|
| Installing and the first commands | Getting Started with the CLI |
| Writing a brief | Briefing Your First AI Agent |
| The gates | The 4-Gate Workflow |
| Why HTML video | Why HyperFrames as the Motion Engine? |
| Keeping a style | Style Memory with TASTE.md |
| Editing without code | Visual Editing in HyperFrames Studio |
| Versions and campaigns | Managing Campaigns vs Version Branches |
| Sound | Loudness Standards · Music Beds · Brand Assets & SFX |
| Other tools | Tracking External Renders · Connecting Agent CLIs |