Skip to main content

Multi-AI MCP Server - Enables Claude Code to collaborate with Gemini, Grok-3, ChatGPT, and DeepSeek

Project description

MseeP.ai Security Assessment Badge

Talk to Multiple AIs Through Claude Code

๐Ÿš€ Make Claude Code even smarter by connecting it to Google Gemini, Grok-3, ChatGPT, and DeepSeek!

โšก 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"

๐Ÿค– Which AIs Are Included?

  • ๐Ÿง  Gemini (Google) - Free API โœ…
  • ๐Ÿš€ Grok-3 (xAI) - Paid API โœ…
  • ๐Ÿ’ฌ ChatGPT (OpenAI) - Paid API โœ…
  • ๐Ÿ”ฎ DeepSeek - Paid API โœ…

๐Ÿ’ก You don't need all of them! Start with just Gemini (it's free), then add others if you want.

๐Ÿš€ 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:

๐Ÿ’ก 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! ๐ŸŽ‰

๐Ÿ’ฌ 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)

๐ŸŽ‰ The beauty is you don't need to remember any commands - just ask Claude naturally!

๐Ÿ”ง 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
  }
}

๐Ÿ’ก 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

๐ŸŒŸ Why Have Multiple AIs?

Think of it like having a team of experts with different personalities:

  • ๐Ÿง  Gemini (Google): The technical expert

    • Great for detailed explanations and accuracy
    • Best for: Complex technical questions, code analysis
  • ๐Ÿš€ Grok (xAI): The creative problem solver

    • Brings unique perspectives and humor
    • Best for: Creative solutions, brainstorming, alternative approaches
  • ๐Ÿ’ฌ ChatGPT (OpenAI): The balanced advisor

    • Comprehensive analysis and practical examples
    • Best for: General advice, code examples, balanced perspectives
  • ๐Ÿ”ฎ 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

๐Ÿ”ง 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.

๐Ÿ’ฐ 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

๐Ÿ”’ 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

โš ๏ธ Important: Never share your credentials.json file or commit it to version control!

๐Ÿ› 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

๐Ÿ“ File Structure

~/.claude-mcp-servers/multi-ai-collab/
โ”œโ”€โ”€ server.py           # Main MCP server
โ”œโ”€โ”€ credentials.json    # API keys and configuration
โ””โ”€โ”€ requirements.txt    # Python dependencies

๐Ÿš€ 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?"

๐Ÿค Contributing

This is designed to be simple and extensible. To add new AI providers:

  1. Add credentials to credentials.json
  2. Add client initialization in server.py
  3. Test with the existing tools

๐Ÿ“œ License

MIT - Use freely!


๐ŸŽ‰ Enjoy having multiple AI assistants working together with Claude Code!

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

File details

Details for the file iflow_mcp_raiansar_claude_code_multi_ai_mcp-1.0.0.tar.gz.

File metadata

  • Download URL: iflow_mcp_raiansar_claude_code_multi_ai_mcp-1.0.0.tar.gz
  • Upload date:
  • Size: 15.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.10.0 {"installer":{"name":"uv","version":"0.10.0","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Debian GNU/Linux","version":"13","id":"trixie","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}

File hashes

Hashes for iflow_mcp_raiansar_claude_code_multi_ai_mcp-1.0.0.tar.gz
Algorithm Hash digest
SHA256 97c532d91cc845aaa44dbc5bc6f24ca9b3946264a06cd37632326748c6a4f882
MD5 6e5e2bfe464547b8fff05bfb9835d5e9
BLAKE2b-256 a0ebf8d4411723a56b8bf3f60cab0af22b33c5754a1ff3552bccc3e4f2c79cd8

See more details on using hashes here.

File details

Details for the file iflow_mcp_raiansar_claude_code_multi_ai_mcp-1.0.0-py3-none-any.whl.

File metadata

  • Download URL: iflow_mcp_raiansar_claude_code_multi_ai_mcp-1.0.0-py3-none-any.whl
  • Upload date:
  • Size: 28.8 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.10.0 {"installer":{"name":"uv","version":"0.10.0","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Debian GNU/Linux","version":"13","id":"trixie","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}

File hashes

Hashes for iflow_mcp_raiansar_claude_code_multi_ai_mcp-1.0.0-py3-none-any.whl
Algorithm Hash digest
SHA256 c9415280c423311aa82c3755e5fe7cde3b737cdd2d61d3a3e6e3239e339e6120
MD5 d6060ee636864c16de43df23d385df98
BLAKE2b-256 679d02de7876d26e13b2ac812db1c45cb5e05d51bee0388bb1f6d21bf40b8c5e

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page