# Kynimo

> Motion graphics for your AI agent. Kynimo is a motion graphics engine for AI assistants. Connect it over MCP, describe a video, and get back editable layers and keyframes, checked against motion guidelines.

Status: in development. The scene format, engine, renderer and MCP server are built. The hosted endpoint, quality report and video export are not available yet. The hosted MCP endpoint is not public yet. It opens with the private beta.

## Start here

- [Landing page](https://web-casio3.vercel.app/): what Kynimo is, what an assistant does with it, how it connects, the style packs and the roadmap
- [Full description for language models](https://web-casio3.vercel.app/llms-full.txt): the whole page as plain text, with the tool list and FAQ

## Source and docs

- [Repository](https://github.com/dalonga-kohi/kynimo): source code, public
- [Product](https://github.com/dalonga-kohi/kynimo/blob/main/PRODUCT.md): what it is and who it is for
- [Architecture](https://github.com/dalonga-kohi/kynimo/blob/main/docs/ARCHITECTURE.md): scene graph, engine, renderer, MCP server
- [Capabilities](https://github.com/dalonga-kohi/kynimo/blob/main/docs/CAPABILITIES.md): what the engine can express
- [Roadmap](https://github.com/dalonga-kohi/kynimo/blob/main/docs/ROADMAP.md): what is built and what is next

## Optional

- [Decision records](https://github.com/dalonga-kohi/kynimo/tree/main/docs/decisions): why each architectural choice was made