Crawl4AI+SearXNG MCP Server

# ======================================== # MCP CRAWL4AI SERVER CONFIGURATION # ======================================== # The transport for the MCP server - either 'sse' or 'stdio' (defaults to sse if left empty) TRANSPORT=sse # Host to bind to if using sse as the transport (leave empty if using stdio) # Set this to 0.0.0.0 if using Docker, otherwise set to localhost (if using uv) HOST=0.0.0.0 # Port to listen on if using sse as the transport (leave empty if using stdio). # Default port for Docker deployment PORT=8051 # Get your Open AI API Key by following these instructions - # https://help.openai.com/en/articles/4936850-where-do-i-find-my-openai-api-key # This is for the embedding model - text-embed-small-3 will be used OPENAI_API_KEY= # The LLM you want to use for summaries and contextual embeddings # Generally this is a very cheap and fast LLM like gpt-4.1-nano-2025-04-14 MODEL_CHOICE=gpt-4.1-nano-2025-04-14 # ======================================== # SEARXNG INTEGRATION CONFIGURATION # ======================================== # SearXNG URL for your search instance - configured for Docker Compose integration # When using Docker Compose, this points to the internal SearXNG service # For external SearXNG instances, change to: https://searx.example.com or http://host.docker.internal:8080 (for docker hosted searxng instances) SEARXNG_URL=http://searxng:8080 # Optional: Custom user agent for SearXNG requests SEARXNG_USER_AGENT=MCP-Crawl4AI-RAG-Server/1.0 # Optional: Default search engines to use (comma-separated list) # Available engines depend on your SearXNG instance configuration SEARXNG_DEFAULT_ENGINES=google,bing,duckduckgo # Optional: Request timeout in seconds for SearXNG API calls SEARXNG_TIMEOUT=30 # ======================================== # OPTIONAL CADDY CONFIGURATION # ======================================== # Optional: Hostname for SearXNG Caddy reverse proxy # Change this to your domain name for production deployment # Default: http://localhost (matches docker-compose.yml default) SEARXNG_HOSTNAME=http://localhost # TLS configuration for Caddy # Use 'internal' for self-signed certificates or provide your email for Let's Encrypt # For production with real domain: your-email@example.com # SEARXNG_TLS=internal # RAG strategies - set these to "true" or "false" (default to "false") # USE_CONTEXTUAL_EMBEDDINGS: Enhances embeddings with contextual information for better retrieval USE_CONTEXTUAL_EMBEDDINGS=false # USE_HYBRID_SEARCH: Combines vector similarity search with keyword search for better results USE_HYBRID_SEARCH=false # USE_AGENTIC_RAG: Enables code example extraction, storage, and specialized code search functionality USE_AGENTIC_RAG=false # USE_RERANKING: Applies cross-encoder reranking to improve search result relevance USE_RERANKING=false # USE_KNOWLEDGE_GRAPH: Enables AI hallucination detection and repository parsing tools using Neo4j # If you set this to true, you must also set the Neo4j environment variables below. USE_KNOWLEDGE_GRAPH=false # For the Supabase version (sample_supabase_agent.py), set your Supabase URL and Service Key. # Get your SUPABASE_URL from the API section of your Supabase project settings - # https://supabase.com/dashboard/project/<your project ID>/settings/api SUPABASE_URL= # Get your SUPABASE_SERVICE_KEY from the API section of your Supabase project settings - # https://supabase.com/dashboard/project/<your project ID>/settings/api # On this page it is called the service_role secret. SUPABASE_SERVICE_KEY= # Neo4j Configuration for Knowledge Graph Tools # These are required for the AI hallucination detection and repository parsing tools # Leave empty to disable knowledge graph functionality # Neo4j connection URI - use bolt://localhost:7687 for local, neo4j:// for cloud instances # IMPORTANT: If running the MCP server through Docker, change localhost to host.docker.internal NEO4J_URI=bolt://localhost:7687 # Neo4j username (usually 'neo4j' for default installations) NEO4J_USER=neo4j # Neo4j password for your database instance NEO4J_PASSWORD=