How it works

From scattered documents to instant AI-powered knowledge in three steps.

1

Upload or Connect

Start by uploading documents directly or connecting a service like Google Drive. Knowledge Raven supports PDF, DOCX, TXT, Markdown, and CSV files.

  • Drag-and-drop file upload in the web interface
  • Google Drive connector with automatic sync
  • Organize documents into knowledge bases and sections
  • Set permissions at the knowledge base or section level
2

Automatic Indexing

Once uploaded, Knowledge Raven automatically processes your documents. No configuration, no embedding knowledge required.

  • Documents are chunked into optimal segments for retrieval
  • Contextual embeddings capture meaning in context
  • Parent-child retrieval: match on small chunks, return full context
  • Hybrid index enables both keyword and semantic search
3

Your Agent Searches

Connect any MCP-compatible LLM (Claude, GPT, Gemini, etc.) and give it the Knowledge Raven skill. Your agent now has specialized tools to search and retrieve from your knowledge base.

  • search_knowledge_base — hybrid search across all documents
  • fetch_document_section — retrieve specific sections with context
  • fetch_full_document — get complete document content
  • get_document_metadata — check document details and structure
  • Every result includes source links for verification