Features
Everything you need to make company knowledge accessible to AI agents — and your team.
Model-Agnostic via MCP
Knowledge Raven uses the Model Context Protocol (MCP) — an open standard for AI tool use. Connect Claude, GPT, Gemini, or any MCP-compatible LLM. Your agent receives specialized tools to search, fetch, and cite from your knowledge base. No vendor lock-in, ever.
Hybrid Search (Vector + BM25)
Combine the precision of keyword search with the semantic understanding of vector search. Our hybrid approach uses Weaviate's native BM25 + vector search to find relevant content even when queries are vague or use different terminology than the source documents.
Multi-Format Support
Upload PDF, DOCX, TXT, Markdown, and CSV files — all automatically chunked, embedded, and indexed. With Gemini Embedding 2's native multimodal capabilities, support for images, audio, and video is on our roadmap.
Smart Connectors
Connect Google Drive and keep your knowledge base in sync automatically. When source documents change, Knowledge Raven detects and re-indexes them. Confluence, Notion, SharePoint, and more connectors are on the roadmap.
Source Deep-Linking
Every piece of information your agent retrieves includes a source link. Click it to open a Reader View that auto-scrolls to the exact cited passage with visual highlighting. Verify any claim in one click.
Admin Dashboard
Understand how your knowledge base is being used. See which documents are searched most, track relevance scores, detect knowledge gaps, and get optimization suggestions — all from a single dashboard.