High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph โ average repo in milliseconds. 158 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.
MCP server for multi-agent development teams
Claude Team is an MCP server that plugs into Claude Code, Cursor, and Windsurf so your agent can coordinate multiple models as a team. It includes task delegation, workflow templates, code review, bug fixing, observability, and collaboration history.
Builders who use Claude Code, Cursor, or Windsurf and want their agent to route work across multiple models and experts.
You can turn one agent session into a multi-model workflow with planning, review, and history instead of handling every step yourself.
What it does
Multi-model collaboration
Coordinates GPT, Claude, Gemini, and custom workers on the same task.
Smart task distribution
Uses a tech-lead step to break work apart and assign it to the best-fit experts.
Workflow templates
Provides prebuilt flows for code generation, bug fixing, refactoring, review, and documentation.
Custom experts
Lets you define experts like Rust, K8s, or Security with environment variables.
Observability tools
Shows team status, cost estimates, and task plans before work runs.
History tools
Keeps collaboration records you can list, inspect, search, and reuse.
How to get it
- 1Run
# Global install npm install -g claude-team # Or use directly with npx (no install needed) npx claude-team
- 2Run
> Help me build a user login feature with the team > Have the team optimize this code for performance
README
๐ค Claude Team
Multi-Agent MCP Server for AI-Powered Development Teams
Orchestrate GPT, Claude, Gemini and more to collaborate on complex tasks
โจ Features
| Feature | Description |
|---|---|
| ๐ค Multi-Model Collaboration | Configure multiple AI models to work together, each leveraging their strengths |
| ๐ง Smart Task Distribution | Tech Lead analyzes tasks and automatically assigns them to the best-suited experts |
| ๐ Workflow Templates | 5 pre-built workflows: code generation, bug fixing, refactoring, review, documentation |
| ๐ฏ Custom Experts | Define your own experts (Rust, K8s, Security, etc.) via environment variables |
| ๐ Observability | Dashboard, cost estimation, and task planning preview |
| ๐ Proxy API Support | Custom Base URLs, compatible with various proxy services |
| ๐ Collaboration History | Complete record of all collaborations with search support |
๐ Quick Start
Installation
# Global install
npm install -g claude-team
# Or use directly with npx (no install needed)
npx claude-team
Basic Configuration
Add to your IDE's MCP configuration file:
๐ Configuration File Locations
| IDE | Path |
|---|---|
| Claude Code | ~/.claude/config.json |
| Windsurf | ~/.codeium/windsurf/mcp_config.json |
| Cursor | ~/.cursor/mcp.json |
{
"mcpServers": {
"claude-team": {
"command": "npx",
"args": ["-y", "claude-team"],
"env": {
"CLAUDE_TEAM_MAIN_KEY": "sk-your-api-key",
"CLAUDE_TEAM_MAIN_URL": "https://api.openai.com/v1",
"CLAUDE_TEAM_MAIN_MODEL": "gpt-4o",
"CLAUDE_TEAM_MAIN_PROVIDER": "openai"
}
}
}
}
Start Using
> Help me build a user login feature with the team
> Have the team optimize this code for performance
๐ฌ How It Works
User: "Optimize this SQL query for performance"
Tech Lead Analysis โ
โโโ Creates: SQL Optimization Expert (powerful)
โโโ Creates: Index Analysis Expert (balanced)
โโโ Workflow: sequential
User: "Build a settings page with dark mode"
Tech Lead Analysis โ
โโโ Creates: UI Component Expert (balanced)
โโโ Creates: Theme System Expert (fast)
โโโ Creates: State Management Expert (balanced)
โโโ Workflow: parallel โ review
๐ ๏ธ Available Tools
Core Tools
| Tool | Description |
|---|---|
team_work | ๐ Team collaboration (auto-creates experts) |
ask_expert | ๐ฌ Consult an expert (supports custom experts) |
code_review | ๐ Code review |
fix_bug | ๐ Bug fixing |
Workflow Tools
| Tool | Description |
|---|---|
list_workflows | ๐ List all workflow templates |
run_workflow | โถ๏ธ Execute a specific workflow |
suggest_workflow | ๐ก Auto-recommend workflow based on task |
Pre-built Workflows:
| Workflow | Purpose | Steps |
|---|---|---|
code-generation | Generate code from requirements | Design โ Implement โ Test โ Review |
bug-fix | Diagnose and fix bugs | Diagnose โ Fix โ Verify |
refactoring | Code refactoring | Analyze โ Plan โ Execute โ Review |
code-review | Multi-dimensional review | Security / Quality / Performance (parallel) |
documentation | Generate documentation | Analyze โ Document |
Observability Tools
| Tool | Description |
|---|---|
team_dashboard | ๐๏ธ View team status, experts, models, stats |
cost_estimate | ๐ฐ Estimate task cost (tokens, price, time) |
explain_plan | ๐ง Preview task assignment plan |
usage_stats | ๐ View model usage statistics |
Integration Tools
| Tool | Description |
|---|---|
read_project_files | ๐ Read project files for context |
analyze_project_structure | ๐๏ธ Analyze project structure and tech stack |
generate_commit_message | ๐ Generate commit message from diff |
History Tools
| Tool | Description |
|---|---|
history_list | ๐ View collaboration history |
history_get | ๐ Get history details |
history_search | ๐ Search history records |
history_context | ๐ Get recent context |
โ๏ธ Configuration
Environment Variables
| Variable | Required | Description |
|---|---|---|
CLAUDE_TEAM_MAIN_KEY | โ | Main model API Key |
CLAUDE_TEAM_MAIN_URL | โ | Main model API URL |
CLAUDE_TEAM_MAIN_MODEL | โ | Main model ID (default: gpt-4o) |
CLAUDE_TEAM_MAIN_PROVIDER | โ | Provider: openai / anthropic / gemini |
CLAUDE_TEAM_MODEL{N}_* | โ | Worker model N config (inherits from MAIN) |
CLAUDE_TEAM_CUSTOM_EXPERTS | โ | Custom experts (JSON format) |
N = 1, 2, 3... supports up to 10 worker models
Custom Experts
Define your own experts beyond the built-in frontend, backend, qa:
{
"env": {
"CLAUDE_TEAM_CUSTOM_EXPERTS": "{\"rust\":{\"name\":\"Rust Expert\",\"prompt\":\"You are a Rust expert...\",\"tier\":\"powerful\"},\"k8s\":{\"name\":\"K8s Expert\",\"prompt\":\"You are a Kubernetes expert...\",\"tier\":\"balanced\"}}"
}
}
| Field | Required | Description |
|---|---|---|
name | โ | Expert display name |
prompt | โ | Expert role description (System Prompt) |
tier | โ | Model tier: fast / balanced / powerful |
skills | โ | Skill tags array |
Model Tiers
| Tier | Use Case | Example Scenarios |
|---|---|---|
fast | Simple, quick tasks | Formatting, simple queries, docs |
balanced | Regular dev tasks | Components, APIs, unit tests |
powerful | Complex reasoning | Architecture, optimization, security |
๐ฆ Changelog
v0.4.0
- ๐ฏ Custom Experts - Define experts via environment variables
- ๐ Workflow Templates - 5 pre-built workflows
- ๐ Observability - Dashboard, cost estimation, plan preview
- ๐ Integration - Project file reading, structure analysis, commit messages
- ๐ก Smart Recommendations - Auto-suggest workflows
- ๐งช Test Coverage - 155 test cases
v0.3.0
- ๐ Task interrupt/resume support
- ๐ฌ Multi-turn expert conversations
- ๐ Token counting and cost estimation
- ๐ Expert templates (6 built-in + custom)
- ๐ Webhook notifications
- โก Exponential backoff retry
- ๐ง Hot config reload
Earlier versions
v0.2.x
- ๐ Streaming output support
- ๐ Usage statistics
- ๐ฏ Model strategies
- ๐พ Result caching
- ๐ Auto model switching
v0.1.x
- ๐ Initial release
- ๐ค Multi-model collaboration
- ๐ Proxy API support
๐ค Contributing
Contributions are welcome! Please read our:
๐ License
Files in the repo
- .github
- src
- .gitignore
- .npmignore
- CHANGELOG.md
- CODE_OF_CONDUCT.md
- CONTRIBUTING.md
- eslint.config.js
- LICENSE
- package-lock.json
- package.json
- README_CN.md
- README.md
- SECURITY.md
- tsconfig.json
Discussion (0)
Ask about usage, or say what you built with itSign in to join the discussion.
No comments yet. Be the first to say what this is good for.
More connectors

Universal provider proxy for OpenAI Codex & Claude Code โ use any LLM (Claude, Gemini, Grok, DeepSeek, Ollamaโฆ) with Codex CLI, App, SDK, and Claude Code
Git-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300ยตs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zero external databases or dependencies. Built in pure Go.
Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.
Stop your AI from making things up โ it proposes, deterministic tools decide, every claim checked against ground truth with evidence. Grounded facts and context survive resets. Reverse engineering is the proving ground. MCP server + CLI.
20 MB lightweight cross-platform database client for 90+ databases, including MySQL, PostgreSQL, SQLite, Redis, MongoDB, DuckDB, SQL Server, and Dameng. Built-in AI, MCP Server, CLI, desktop and Docker. | ่ฝป้็บง่ทจๅนณๅฐๆฐๆฎๅบ็ฎก็ๅทฅๅ ท๏ผๆฏๆ MySQLใPostgreSQLใSQLiteใRedisใMongoDBใ่พพๆขฆ็ญ 90+ ๆฐๆฎๅบ๏ผๆไพๆก้ข็ซฏใDockerใCLIใๅ ็ฝฎ AI ๅฉๆๅ MCP Serverใ