lightfast-mcp

# Examples Directory This directory contains example scripts and demonstrations for the Lightfast MCP modular architecture. ## Demo Scripts ### `demo_modular_system.py` Comprehensive demonstration of the complete modular MCP server architecture: - Server auto-discovery - Multi-server management - AI client integration with multiple servers - Configuration management - Health monitoring and cleanup **Usage**: `uv run python examples/demo_modular_system.py` ### `demo_ai_integration.py` Demo showing AI integration workflow with Blender MCP server without requiring API keys: - Basic MCP server connection testing - Tool execution simulation - AI workflow simulation - Setup guide for real AI integration **Usage**: `uv run python examples/demo_ai_integration.py` ## Reference Implementations ### `ai_blender_client.py` Simple single-server AI client specifically for Blender MCP integration: - Direct connection to one Blender MCP server - AI conversation with tool execution - Example of using official Anthropic/OpenAI SDKs **Usage**: `uv run python examples/ai_blender_client.py` **Note**: For production use, consider the multi-server orchestrator: `uv run lightfast-mcp-orchestrator ai` ## Requirements - Make sure you have the modular system installed: `uv install` - For AI integration examples, set environment variables: - `ANTHROPIC_API_KEY` for Claude - `OPENAI_API_KEY` for OpenAI - `AI_PROVIDER` (optional, defaults to "claude") ## Related Scripts - Main multi-server orchestrator: `uv run lightfast-mcp-orchestrator` - Test system: `uv run python scripts/test_working_system.py` - Blender-specific testing: `./scripts/test_blender.sh`