What it does
cclsp bridges Language Server Protocol (LSP) and Model Context Protocol (MCP), enabling AI-powered coding agents to reliably navigate and manipulate codebases. It solves a core integration challenge: AI assistants often produce imprecise line and column numbers, which breaks naive LSP integration. cclsp handles this by intelligently trying multiple position combinations to find the intended target. It exposes tools for finding definitions, locating all references to a symbol, renaming symbols safely, and retrieving diagnostics from language servers.
Who it's for
LLM-based coding agents—particularly Claude Code—working in non-IDE environments or integrated into custom development workflows. Also useful for developers who want to leverage their language server investment (TypeScript, Python, Go, Rust, etc.) through an MCP interface, without being tied to a specific IDE.
Common use cases
- Jump to a function or class definition within a codebase
- Find all references to a symbol across files
- Rename a symbol safely throughout an entire codebase
- Retrieve linting and type-checking diagnostics
- Navigate unfamiliar code while working with an LLM assistant
Setup pitfalls
- Requires Node.js 18+ or Bun runtime
- Language servers must be installed separately for each language—
npm install -g typescript-language-server,pip install python-lsp-server,go install golang.org/x/tools/gopls@latest, etc. - Configuration file must be created at
.claude/cclsp.jsonor~/.config/claude/cclsp.json; the setup wizard can automate this - Requires read and write access to the filesystem to interact with LSPs