multiple-model-mcp

Multiple Model MCP: 17+ FREE AI Models in One Server

The ultimate multi-model MCP server for Claude - gives you access to 17+ completely FREE AI models including Llama 4, DeepSeek R1, Qwen 3, Microsoft Phi-4, and more for enhanced code analysis, problem-solving, and collaborative development.

Features true AI orchestration with conversations that continue across tasks - Give Claude a complex
task and let it orchestrate between the best FREE models automatically. Claude stays in control, performs the actual work,
but gets perspectives from the best AI for each subtask. With tools like analyze for
understanding codebases, codereview for audits, refactor for
improving code structure, debug for solving complex problems, and precommit for
validating changes, Claude can switch between different tools and models mid-conversation,
with context carrying forward seamlessly.

🆓 FREE Models Available (17+)

Meta Llama Models (Latest!)

• llama-4-scout - Latest Llama 4 model (FREE!)
• llama-4-maverick - Latest Llama 4 variant (FREE!)
• llama-3.1-nemotron-ultra-253b - Massive 253B parameter model

DeepSeek Models (Best for Reasoning)

• deepseek-r1-0528 - Latest R1 reasoning model
• deepseek-r1-distill-llama-70b - Groq ultra-fast reasoning

Qwen Models (Alibaba's Latest)

• qwen3-235b - Massive 235B parameter model (FREE!)
• qwen3-32b - Balanced performance

Microsoft Models

• phi-4-reasoning-plus - Enhanced Phi-4 reasoning
• mai-ds-r1 - Latest MAI model

Google Gemini Models

• flash - Fast Gemini Flash (FREE tier!)
• gemini-2.0-flash-exp - Experimental version

Other Top FREE Models

• mistral-small-3.1-24b-instruct - Mistral's latest
• gemma2-9b-it - Google Gemma 2
• mixtral-8x7b-32768 - Mixture of experts
• glm-z1-32b - GLM reasoning model

Example Workflow - Claude Code:

1. Performs its own reasoning
2. Uses Llama 4 Scout to deeply analyze the code for a second opinion
3. Switches to DeepSeek R1 to continue chatting about reasoning
4. Uses Qwen 3 235B to evaluate complex logic from DeepSeek
5. Performs the actual work after taking in feedback from all models
6. Returns to Phi-4 for a precommit review

All within a single conversation thread! Llama 4 in step 6 knows what was recommended by DeepSeek in step 3!

🚀 Quick Start

Option 1: Smithery (Recommended)

The easiest way to get started with all 17+ FREE models:

1. Install Smithery
2. Add Multiple Model MCP Server:

   smithery install samihalawa/multiple-model-mcp
   

3. Configure your FREE models and start using!

Option 2: Docker (Traditional)

git clone https://github.com/samihalawa/multiple-model-mcp.git
cd multiple-model-mcp
./run-server.sh

Option 3: Claude Code CLI

claude mcp add multiple-model-mcp -s user -- docker exec -i multiple-model-mcp python server.py

🛠️ Available Tools

1. Chat - General Development Chat & Collaborative Thinking

Perfect for brainstorming, asking questions, or getting a second opinion on complex problems.

Best Models: llama-4-scout, deepseek-r1-0528, qwen3-235b

{
  "name": "chat",
  "arguments": {
    "message": "How should I implement caching for this API?",
    "model": "llama-4-scout"
  }
}

2. ThinkDeep - Extended Reasoning & Complex Problem Solving

For problems requiring deep analysis and extended reasoning chains.

Best Models: deepseek-r1-0528, phi-4-reasoning-plus

{
  "name": "thinkdeep", 
  "arguments": {
    "task": "Design a scalable microservices architecture",
    "model": "deepseek-r1-0528",
    "thinking_mode": "high"
  }
}

3. CodeReview - Professional Code Review

Comprehensive code analysis with security, performance, and best practice recommendations.

Best Models: qwen3-235b, llama-4-maverick

{
  "name": "codereview",
  "arguments": {
    "files": ["src/api.py", "src/models.py"],
    "model": "qwen3-235b"
  }
}

4. PreCommit - Pre-commit Validation

Fast validation before commits with focus on immediate issues.

Best Models: gemma2-9b-it, mistral-small-3.1-24b-instruct

{
  "name": "precommit",
  "arguments": {
    "files": ["changed_file.py"],
    "model": "gemma2-9b-it"
  }
}

5. Debug - Expert Debugging Assistant

Systematic debugging approach for complex issues.

Best Models: deepseek-r1-0528, llama-4-scout

{
  "name": "debug",
  "arguments": {
    "issue": "Memory leak in long-running process",
    "files": ["src/processor.py"],
    "model": "deepseek-r1-0528"
  }
}

6. Analyze - Smart File Analysis

Deep understanding of code structure, patterns, and relationships.

Best Models: qwen3-235b, llama-4-maverick

{
  "name": "analyze",
  "arguments": {
    "files": ["entire_codebase/"],
    "focus": "architecture patterns",
    "model": "qwen3-235b"
  }
}

7. Refactor - Intelligent Code Refactoring

Code improvements while preserving functionality.

Best Models: llama-4-scout, mixtral-8x7b-32768

{
  "name": "refactor",
  "arguments": {
    "files": ["legacy_code.py"],
    "goal": "improve maintainability",
    "model": "llama-4-scout"
  }
}

8. TestGen - Smart Test Generation

Generate comprehensive test suites with edge cases.

Best Models: phi-4-reasoning-plus, deepseek-r1-0528

{
  "name": "testgen",
  "arguments": {
    "files": ["src/utils.py"],
    "test_types": ["unit", "integration"],
    "model": "phi-4-reasoning-plus"
  }
}

9. Tracer - Execution Flow Analysis

Understand code execution paths and data flow.

Best Models: qwen3-235b, deepseek-r1-distill-llama-70b

{
  "name": "tracer",
  "arguments": {
    "files": ["complex_algorithm.py"],
    "model": "qwen3-235b"
  }
}

🔧 Configuration

All models are FREE to use! Simply configure the providers you want:

• OpenRouter: Access to all FREE models via single API
• Gemini: Direct Google API for Flash models
• Groq: Ultra-fast inference for Llama, DeepSeek, Mixtral
• Custom APIs: Support for local models (Ollama, vLLM, etc.)

🤝 Contributing

Contributions welcome! This server provides access to the best FREE AI models available.

📄 License

MIT License - Use these FREE models however you want!

🔗 Links

• Repository: https://github.com/samihalawa/multiple-model-mcp
• Smithery Package: Available on Smithery.ai
• Documentation: See /docs folder for detailed guides

---

Multiple Model MCP Server - Because why pay for AI when you can get 17+ world-class models for FREE? 🆓