-
Notifications
You must be signed in to change notification settings - Fork 15
Description
Hey team!
I've been using the Kill Bill docs extensively for some time now, and think it would be awesome if you added an MCP (Model Context Protocol) server for the Slate documentation. This would let AI assistants like Claude directly access and search through your docs.
What's an MCP server?
It's basically a standardized way for AI tools to connect to data sources. Instead of AI assistants having to scrape web pages or rely on outdated training data, they could query your docs directly through a proper API.
Why this would be useful:
- Developers could ask their AI assistant questions about Kill Bill and get accurate, up-to-date answers pulled straight from your docs
- No more outdated info from LLM training data
- Makes your docs more accessible to the growing number of people using AI coding assistants
- Allows AI assistants to piece together custom flows for testing purposes
Our team uses an MCP server for our own API documentation, and has had great success with building some powerful automation for various tasks from it.
Possible Implementation:
The MCP TypeScript SDK (https://github.com/modelcontextprotocol/typescript-sdk) makes this pretty straightforward. You'd essentially create a server that exposes your Slate docs content through the MCP protocol.
Note about Slate:
I checked and Slate itself doesn't have built-in MCP support as far as I can tell (it's focused on static site generation), so this would need to be a custom MCP server that reads from your Slate markdown files or the generated site.