
Multi-AI MCP Server for Claude Code
Connects Claude Code with multiple AI models (Gemini, Grok-3, ChatGPT, DeepSeek) simultaneously, allowing users to get diverse AI perspectives, conduct AI debates, and leverage each model's unique strengths.
README
Talk to Multiple AIs Through Claude Code
🚀 Make Claude Code even smarter by connecting it to Google Gemini, Grok-3, ChatGPT, and DeepSeek!
⚡ You can use any combination - just the ones you have API keys for!
Instead of switching between different AI websites, now you can simply ask Claude Code to get help from other AIs. Just say things like:
"Hey Claude, ask all the AIs to help debug this code"
"Claude, get Grok's opinion on this architecture"
"Have Gemini and ChatGPT debate this technical decision"
🤖 Which AIs Are Included?
- 🧠 Gemini (Google) - Free API ✅
- 🚀 Grok-3 (xAI) - Paid API ✅
- 💬 ChatGPT (OpenAI) - Paid API ✅
- 🔮 DeepSeek - Paid API ✅
💡 You don't need all of them! Start with just Gemini (it's free), then add others if you want.
🚀 5-Minute Setup
What You Need
- Claude Code installed
- At least one API key (Gemini is free!)
Installation (Copy & Paste These 3 Commands)
# 1. Download the code
git clone https://github.com/RaiAnsar/claude_code-multi-AI-MCP.git
cd claude_code-multi-AI-MCP
# 2. Run the automatic setup
chmod +x setup.sh
./setup.sh
# 3. That's it! Start using it right away
During setup, you'll be asked for API keys:
- Gemini (Free): Get key here
- Grok (Paid): Get key here
- OpenAI (Paid): Get key here
- DeepSeek (Paid): Get key here
💡 Pro tip: Start with just Gemini (it's free), then add others later if you want.
Test It Works
After setup, just ask Claude naturally:
"Hey Claude, ask Gemini what the capital of France is"
If you see a response from Gemini, you're all set! 🎉
💬 How to Use It (Super Simple!)
Once installed, you just talk to Claude Code normally and ask it to use the other AIs. Here are real examples:
Ask Claude to Get Multiple Opinions
You: "Hey Claude, ask all the AIs what they think about using microservices vs monolith architecture"
Claude: I'll ask all available AIs for their perspectives on this...
(Claude will use the
ask_all_ais
tool and show you all available AI responses)
Get Help Debugging Code
You: "Claude, can you have Grok help debug this Python function that's running slowly?"
Claude: Let me ask Grok to analyze your code for performance issues...
(Claude will use the
grok_debug
tool)
Compare Different AI Opinions
You: "Have Gemini and ChatGPT debate whether to use React or Vue for my frontend"
Claude: I'll set up a debate between Gemini and ChatGPT on this topic...
(Claude will use the
ai_debate
tool)
Get Code Reviews from Multiple AIs
You: "Can you ask all the AIs to review this authentication function for security issues?"
Claude: I'll have all available AIs review your code...
(Claude will use multiple code_review tools)
Brainstorm Creative Solutions
You: "Ask Grok to brainstorm some creative features for my todo app"
Claude: Let me get Grok's creative input on your todo app...
(Claude will use the
grok_brainstorm
tool)
Get Architecture Advice
You: "Claude, have Gemini help design the database schema for my e-commerce site"
Claude: I'll ask Gemini to provide architecture recommendations...
(Claude will use the
gemini_architecture
tool)
🎉 The beauty is you don't need to remember any commands - just ask Claude naturally!
🔧 Configuration
API Keys & Models
Edit ~/.claude-mcp-servers/multi-ai-collab/credentials.json
:
{
"gemini": {
"api_key": "your-gemini-key",
"model": "gemini-2.0-flash",
"enabled": true
},
"grok": {
"api_key": "your-grok-key",
"model": "grok-3",
"enabled": true
},
"openai": {
"api_key": "your-openai-key",
"model": "gpt-4o",
"enabled": true
}
}
💡 Pro Tip: You can mix and match any combination:
- Only Gemini? Works perfectly!
- Gemini + ChatGPT? Great for comparing Google vs OpenAI perspectives!
- Any combination? Maximum AI collaboration power!
Getting API Keys
- Gemini: Google AI Studio (Free)
- Grok: xAI Console (Paid)
- OpenAI: OpenAI Platform (Paid)
- DeepSeek: DeepSeek Platform (Paid)
🌟 Why Have Multiple AIs?
Think of it like having a team of experts with different personalities:
-
🧠 Gemini (Google): The technical expert
- Great for detailed explanations and accuracy
- Best for: Complex technical questions, code analysis
-
🚀 Grok (xAI): The creative problem solver
- Brings unique perspectives and humor
- Best for: Creative solutions, brainstorming, alternative approaches
-
💬 ChatGPT (OpenAI): The balanced advisor
- Comprehensive analysis and practical examples
- Best for: General advice, code examples, balanced perspectives
-
🔮 DeepSeek: The reasoning specialist
- Strong in math, coding, and logical reasoning
- Best for: Complex algorithms, mathematical problems, code optimization
Real Benefits:
- Better Decisions: Get 2-3 opinions before making important choices
- Learn Faster: See how different AIs approach the same problem
- Catch Mistakes: If one AI misses something, another might catch it
- Save Time: Get multiple expert opinions without switching apps
🔧 Partial Configurations
Don't have all the API keys? No problem!
- Only Gemini? You'll have access to Google's powerful free AI
- Only Grok? Get xAI's unique perspective and humor
- Only ChatGPT? Use OpenAI's well-established models
- Only DeepSeek? Get specialized reasoning and coding help
- Gemini + ChatGPT? Compare Google vs OpenAI approaches!
- Grok + DeepSeek? Creative thinking meets logical reasoning!
- Any 3 AIs? Excellent multi-perspective collaboration!
- Have all 4? Ultimate AI collaboration with maximum diversity!
The server automatically adapts to your available AIs. Tools for unavailable AIs simply won't appear in Claude Code.
💰 Cost-Effective Options:
- Start with free Gemini to test the system
- Add ChatGPT for proven OpenAI capabilities
- Include Grok for unique xAI insights
- Add DeepSeek for specialized reasoning tasks
🔒 Security Notes
- API keys stored locally: All credentials are in
~/.claude-mcp-servers/multi-ai-collab/credentials.json
- Never committed to git: The
.gitignore
file excludes all credential files - Optional AIs: Only AIs with valid keys are loaded
- Graceful failures: Failed AI connections don't break the server
- Input validation: All requests are validated before processing
⚠️ Important: Never share your credentials.json
file or commit it to version control!
🐛 Troubleshooting
MCP not showing up?
claude mcp list
# Should show "multi-ai-collab"
AI not responding?
mcp__multi-ai-collab__server_status
# Check which AIs are enabled and working
Connection errors?
- Verify API keys in
credentials.json
- Check if the AI service is down
- Ensure you have sufficient API credits
Reinstall:
claude mcp remove multi-ai-collab
./setup.sh
📁 File Structure
~/.claude-mcp-servers/multi-ai-collab/
├── server.py # Main MCP server
├── credentials.json # API keys and configuration
└── requirements.txt # Python dependencies
🚀 Advanced Usage
Temperature Control for All AIs
# Creative writing with high temperature
mcp__multi-ai-collab__ask_gemini
prompt: "Write a creative story about AI collaboration"
temperature: 0.9
mcp__multi-ai-collab__ask_openai
prompt: "Write a creative story about AI collaboration"
temperature: 0.9
# Technical explanations with low temperature
mcp__multi-ai-collab__ask_grok
prompt: "Explain how TCP/IP works"
temperature: 0.2
Specialized Code Reviews by AI
# Gemini: Technical accuracy focus
mcp__multi-ai-collab__gemini_code_review
code: "[your code]"
focus: "technical accuracy"
# ChatGPT: Best practices focus
mcp__multi-ai-collab__openai_code_review
code: "[your code]"
focus: "best practices"
# Grok: Creative solutions focus
mcp__multi-ai-collab__grok_code_review
code: "[your code]"
focus: "alternative approaches"
Multi-AI Workflows
# Step 1: Get all perspectives
mcp__multi-ai-collab__ask_all_ais
prompt: "How should I structure a microservices architecture?"
# Step 2: Have top AIs debate specifics
mcp__multi-ai-collab__ai_debate
topic: "Event-driven vs REST for microservices communication"
ai1: "gemini"
ai2: "openai"
# Step 3: Get Grok's creative alternative
mcp__multi-ai-collab__ask_grok
prompt: "What's a creative alternative to traditional microservices?"
🤝 Contributing
This is designed to be simple and extensible. To add new AI providers:
- Add credentials to
credentials.json
- Add client initialization in
server.py
- Test with the existing tools
📜 License
MIT - Use freely!
🎉 Enjoy having multiple AI assistants working together with Claude Code!
推荐服务器

Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。