Knowledge Base
MCPs for Notion, Confluence, and documentation
TL;DR
Knowledge base MCP servers expose wiki and docs platforms — Notion, Confluence, Obsidian, GitBook — to agents. They power internal Q&A bots, keep docs in sync with code, and let agents draft/edit pages naturally. Core to any RAG-powered internal search.
About Knowledge Base
Knowledge base MCPs connect agents to Notion, Confluence, Obsidian, GitBook, and Mintlify. Agents can read docs, draft pages, answer from internal knowledge, and keep documentation in sync with code.
Common use cases
- Build an internal Q&A bot over your Notion workspace
- Keep API docs in sync with code via agent-authored updates
- Summarize long docs into TL;DR pages
- Auto-generate onboarding docs for new employees
- Index PDFs + web pages into a unified internal knowledge base
MCPs tagged “Knowledge Base”
Related recipes
Invoice Auto-Documentation
Every Stripe invoice is automatically documented in a Notion database with amount, customer, and payment status.
Text-to-Speech Content Generation
Notion pages tagged as 'podcast' are automatically converted to audio files via ElevenLabs and saved back as attachments.
Research Automation
Paste a research topic in Notion and an agent uses Perplexity to gather sources, summarize findings, and structure them.
Competitor Watch Automation
Run daily Perplexity searches on competitors and log product updates, pricing changes, and news to a Notion tracker.
Frequently asked questions
Which KB MCP has the best search?
Notion's official MCP has strong semantic search. Confluence is catching up. For pure search, combine any KB MCP with Meilisearch or Algolia.
Can agents edit wiki pages?
Yes — Notion, Confluence, and Obsidian MCPs all expose `update_page` and `create_page` tools. Always gate destructive edits with approval.
How fresh is the indexed content?
Official MCPs query live — no stale indexes. If you use a secondary search layer (Meilisearch), set a re-index cron (hourly/daily).
Can I use a KB MCP for RAG?
Yes — fetch via the KB MCP, embed with a vector MCP (Pinecone/pgvector), and let the agent retrieve the top-k chunks per query.
Install Knowledge Base MCPs
Browse the full directory or explore all tags to find the right MCPs for your stack.