AI coding assistants that actually know our SDKs
AI coding assistants like GitHub Copilot and Cursor are great at general JavaScript, but they struggle with specialized SDKs. Our MCP (Model Context Protocol) servers solve this by giving these AI tools direct access to our documentation, type definitions, and examples.
Watch Ivan Krushkov demo how an AI assistant can query ReGraph documentation in real-time, understand our SDK’s structure, and generate working code that follows our best practices.
Get working code, not hallucinated API calls
When you ask your AI assistant to build a graph visualization, it can query the actual KeyLines, ReGraph, KronoGraph, or MapWeave documentation and generate code that works. No more guessing, no more debugging AI-generated code that doesn’t match our API.
Use the MCP server with the SDK that fits your application:
FAQ
What is MCP?
Model Context Protocol is an open standard (like USB-C for AI) that lets AI coding assistants connect to external documentation and tools. Think of it as giving your AI assistant the ability to “read the manual” while helping you code.
Do I need access to more than one SDK?
No, each SDK has its own MCP server. Use the one(s) relevant to your project — KeyLines for graphs, ReGraph for React-based graphs, KronoGraph for timelines, MapWeave for geospatial visualization.
Which tools does this work with?
Any IDE or AI assistant that supports MCP — including VS Code with GitHub Copilot, Cursor, JetBrains, Claude Code, and more. As long as your tool can connect to MCP servers, it will work.
Is my code and data secure?
MCP servers only provide read access to our public documentation. Your code stays in your development environment. The MCP protocol lets your AI assistant query our documentation — no proprietary code or data is sent to our servers.
