Chatify

Inspiration

When I first discovered MCP Apps, I got one of the UIs that I made running on her ChatGPT and she was absolutelty blown away. It really stuck with me. I realized how much joy something like this could bring to others. I wanted to make that experience accessible to anyone: a way to create a personalized UI inside their AI client of choice, with no coding or setup: just an idea.

What it does

Chatify takes a plain-English description of any UI idea and returns a single URL that renders your custom interface live inside Claude, ChatGPT, or any MCP-compatible AI client.

How we built it

An AI agent writes a FastMCP Python server from your description and deploys it instantly to Modal, which hosts it as a persistent, publicly-accessible web endpoint with all dependencies pre-baked into a cached base image.

Challenges we ran into

The biggest challenge was making AI-generated FastMCP servers reliable enough to deploy and run correctly on the first attempt, every time, without any human debugging.

Accomplishments that we're proud of

We built a pipeline that takes a user from raw idea to live, running MCP app in seconds — entirely autonomously, with no code ever touched by the user.

What we learned

Prompt engineering is infrastructure — the reliability of the entire platform lives or dies on how well the system prompt guides the agent to write valid, deployable FastMCP code.

What's next for Chatify

We plan to add persistent user data across sessions, expand the UI component library available to the agent, and build a gallery where users can share and remix each other's Chatify creations.

Built With

• fastmcp
• mcp-apps
• modal
• python