mcpls and lsp-mcp-server
These tools are competitors, as both aim to provide a universal bridge between MCP and LSP, with A translating Claude Code to LSP and B exposing LSP capabilities as MCP tools for AI agents.
About mcpls
bug-ops/mcpls
Universal MCP to LSP bridge - expose Language Server Protocol capabilities as MCP tools for AI agents
Bridges MCP and LSP by spawning language servers (rust-analyzer, pyright, gopls, etc.) and exposing their capabilities—type inference, cross-references, diagnostics, call hierarchies—as MCP tools. Uses project marker heuristics to automatically detect and start only relevant servers, with graceful degradation if servers fail. Integrates with AI coding assistants like Claude through MCP's stdio transport, enabling semantic code understanding beyond text-based reasoning.
About lsp-mcp-server
ProfessioneIT/lsp-mcp-server
An MCP (Model Context Protocol) server that bridges Claude Code to Language Server Protocol (LSP) servers, enabling semantic code intelligence capabilities.
Exposes 24 MCP tools providing semantic code intelligence (go-to-definition, find references, rename, diagnostics, call/type hierarchy) across 10 languages via stdio-based JSON-RPC communication with language servers. Features automatic server lifecycle management, multi-root workspace support with per-workspace server instances, and push-based diagnostics caching for real-time error tracking. Integrates with any LSP-compatible server (TypeScript, Python, Rust, Go, C/C++, Ruby, PHP, Elixir, Kotlin, Java) and configures directly into Claude Code via `.mcp.json`.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work