kimi-code-mcp
Active·★ 61·MIT·Updated 2026-10-12
This MCP server integrates Kimi Code (K2.5) with Claude Code, allowing Claude to delegate bulk code reading and analysis to Kimi while focusing on decision-making and code editing. It reduces Claude token consumption by using Kimi's 256K context for cheap, deep code understanding. The tool supports code analysis, querying, session management, and can be configured for different levels of output detail.
#ai-agent#ai-coding-assistant#claude-code#code-analysis#developer-tools#kimi#kimi-code#kimi-k2
01
Features
01Delegates bulk code reading to Kimi for token savings
02256K context for deep codebase understanding
03Provides structured analysis (architecture, security, dependencies)
04Resume sessions with retained context
05Configurable output detail levels (summary/normal/detailed)
02
Compatibility
macOS
macOS
Verified via docs
Linux
Linux
Verified via docs
Windows
Windows
Verified via docs
03
Quick start
1
$ curl -L code.kimi.com/install.sh | bash
2
$ npm install -g kimi-mcp-server
04
Use cases
↳Analyze codebase architecture: Kimi reads all files, Claude acts on the report
↳Scan for security vulnerabilities: Kimi audits, Claude cross-examines
↳Dependency mapping and refactoring planning: Kimi builds graph, Claude plans changes
05
Alternatives
gograph★ 230
A fast, local-only CLI tool to generate repository structures and improve IDE context awareness for Go codebases.
qveris-agent-toolkit★ 262
Open-source toolkit for the QVeris capability routing network: CLI, MCP server, Python SDK, skills, and REST API docs for agents to discover, inspect, call, and audit real-world tools.
codex-tldraw-mcp★ 25
Local Codex MCP server for scanning repositories and writing tldraw product workflow diagrams.
delimit-mcp-server★ 23
Unify Claude Code, Codex, Cursor, and Gemini CLI with persistent context, governance, and multi-model debate.
claude-soul★ 92
Self-correcting learning engine for Claude Code — persistent identity, behavioral pattern tracking, and cross-session memory
Comments
Log in to leave a comment
No comments yet. Be the first!