PDF Knowledgebase MCP Server

# Search Features Guide This guide is being expanded. For now: - Defaults: parser = `pymupdf4llm`; embeddings provider = `local` (Qwen/Qwen3-Embedding-0.6B); chunker = `langchain`; hybrid search = enabled; reranker = disabled; summarizer = disabled. - See the [Quick Start](quick-start.md), [Troubleshooting](troubleshooting.md), and [Docker Deployment](docker-deployment.md) guides for working configurations. For immediate help, see: - [Quick Start Guide](quick-start.md) for basic search setup - [Troubleshooting Guide](troubleshooting.md) for search quality issues This guide will include: - Hybrid search configuration - Reranking with local and DeepInfra models - Semantic chunking strategies - Search optimization techniques **Coming soon...**