MCP Ollama Consult Server

# MCP Ollama Consult - Quick Start Guide Get up and running with MCP Ollama Consult in under 5 minutes! ## Prerequisites - Node.js 18+ - Ollama installed and running locally - An MCP-compatible client (e.g., Claude Desktop) ## 1. Installation ```bash # Clone and setup git clone https://github.com/Atomic-Germ/mcp-consult.git cd mcp-consult npm install npm run build ``` ## 2. Verify Ollama ```bash # Check Ollama is running curl http://localhost:11434/api/tags # Should return list of available models # If not running: ollama serve ``` ## 3. Configure MCP Client ### Claude Desktop Edit your configuration file: - **macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json` - **Windows**: `%APPDATA%\Claude\claude_desktop_config.json` - **Linux**: `~/.config/claude/claude_desktop_config.json` Add the server: ```json { "mcpServers": { "consult": { "command": "node", "args": ["/absolute/path/to/mcp-consult/dist/index.js"], "env": { "OLLAMA_BASE_URL": "http://localhost:11434" } } } } ``` ### Other MCP Clients Use these connection details: - **Command**: `node` - **Args**: `["/path/to/mcp-consult/dist/index.js"]` - **Protocol**: stdio ## 4. Test the Setup ### Option A: Demo Client ```bash npm run demo ``` ### Option B: Claude Desktop 1. Restart Claude Desktop 2. Look for "consult" in the MCP tools sidebar 3. Try asking: "List available Ollama models" ## 5. First Consultation Try these example prompts in Claude: ### Basic Consultation ``` Use the consult_ollama tool to ask llama3.2: "What are the benefits of microservices architecture?" ``` ### Model Comparison ``` Use compare_ollama_models to ask these models the same question: "Explain dependency injection in simple terms" Models: ["llama3.2", "qwen2.5-coder:7b"] ``` ### Sequential Chain ``` Use sequential_consultation_chain with these consultants: 1. analyst (llama3.2): "Analyze the pros and cons of TypeScript" 2. architect (qwen2.5-coder:7b): "Based on the analysis from {analyst}, recommend best practices for a new TypeScript project" ``` ## 6. Verify Everything Works ✅ **Models Listed**: You can see available Ollama models ✅ **Basic Consultation**: Single model responses work ✅ **Comparison**: Multiple models respond to same prompt ✅ **Chains**: Sequential consultations with references ✅ **Memory**: Consultation results can be stored/retrieved ## Common Issues ### "Connection refused" errors - Ensure Ollama is running: `ollama serve` - Check the base URL: `OLLAMA_BASE_URL=http://localhost:11434` ### "No models available" - Pull a model: `ollama pull llama3.2` - Verify: `ollama list` ### MCP client can't find tools - Check the absolute path in configuration - Restart your MCP client after config changes - Verify build succeeded: `ls -la dist/` ## Next Steps - 📖 **Full Documentation**: See [README.md](./README.md) - 🔗 **Sequential Chains**: Check [sequential_chain_demos.md](./sequential_chain_demos.md) - 🏗️ **Architecture**: Read [ARCHITECTURE.md](./ARCHITECTURE.md) - 🤝 **Contributing**: See [CONTRIBUTING.md](./CONTRIBUTING.md) ## Advanced Configuration ### Memory Setup For persistent memory across sessions: ```bash # Use file storage export MEMORY_DIR="/path/to/memory" # Or connect to memory MCP server export REMEMBER_MCP_CONFIG='{"type":"stdio","command":"node","args":["memory-server.js"]}' ``` ### Custom Ollama Endpoint ```bash # Remote Ollama instance export OLLAMA_BASE_URL="http://your-server:11434" ``` ### Timeout Configuration For complex reasoning tasks, increase timeouts in chain consultants: ```json { "consultants": [ { "id": "deep-thinker", "model": "llama3.2", "prompt": "Analyze this complex problem...", "timeoutMs": 300000 } ] } ``` --- **🎉 You're ready to start consulting with Ollama models via MCP!** Need help? Check the [troubleshooting guide](./docs/TROUBLESHOOTING.md) or [open an issue](https://github.com/Atomic-Germ/mcp-consult/issues).