Mcp Pyright
Give Claude Code or Cursor Pyright diagnostics, type info, and language-server features while you edit Python in the repo.
Overview
mcp-pyright is a MCP server for the Build phase that exposes Pyright language-server diagnostics and type intelligence to AI coding agents over stdio.
What is this MCP server?
- Stdio MCP server packaging Pyright as agent-callable language-server behavior
- PyPI package mcp-pyright at version 0.1.0 with stdio transport
- Surfaces type checking and IDE-style Python intelligence inside MCP-compatible clients
- Repository: github.com/daedalus/mcp-pyright
- No cloud API key required for local Pyright runs beyond a Python environment
- Server version 0.1.0
- PyPI identifier mcp-pyright
- Transport: stdio
What problem does it solve?
Agents edit Python blind to Pyright’s view of your types and config unless you paste errors or rerun checks by hand.
Who is it for?
Indie builders using Claude Code or Cursor on typed Python codebases who want LSP-grade feedback inside the agent.
Skip if: Teams that only touch JavaScript/TypeScript or who already get sufficient analysis from built-in IDE tooling without MCP.
What do I get? / Deliverables
Your agent can query Pyright-backed analysis in-session so fixes align with strict typing and pyrightconfig before you ship.
- Registered stdio MCP server entry for Pyright-backed tools
- In-session type and diagnostic context for Python files in the workspace
- Fewer untyped regressions before local test or ship checks
Recommended MCP Servers
Journey fit
Canonical shelf is Build because the value is wiring editor-grade Python analysis into the agent loop during implementation, not pre-build research or post-ship ops. agent-tooling fits an MCP stdio bridge that exposes LSP capabilities to coding agents rather than shipping a standalone app feature.
How it compares
MCP bridge to Pyright LSP, not a standalone Claude skill or a hosted CI service.
Common Questions / FAQ
Who is mcp-pyright for?
Solo and small-team developers who ship Python with Pyright and want the same analysis surfaced to Claude Code, Cursor, or other MCP clients.
When should I use mcp-pyright?
Use it during Build when you are iterating on Python modules, refactoring typed APIs, or debugging type errors the agent introduced.
How do I add mcp-pyright to my agent?
Install mcp-pyright from PyPI, add a stdio MCP server entry pointing at that command in Claude Desktop, Claude Code, or Cursor, and open the project root your pyrightconfig targets.