AI Native Knowledge Base
Structured pages, sidecar annotations, and an MCP server — so your agents can manage your knowledge base the same way your team does.
Docs rot. Context gets lost. Nobody notices until something breaks.
Ghost ownership
The onboarding guide was last verified by someone who left in Q2.
Version drift
Your API docs say v2 but you shipped v4 — a customer files a ticket before anyone notices.
Lost context
Someone mentions a competitive threat on a call. It lives in Granola notes and nowhere else. The customer page never gets updated.
Agent tax
Your AI agent burns tokens running find and grep trying to figure out where content lives.
Sound familiar?
Annotations close the loop
Drop context from a call, a Slack thread, or your own notes — right onto the page that matters. Annotations live as sidecar files, not embedded in content. When an agent refreshes the page, it reads your annotations first and incorporates what's relevant.
See it in action
Built for your role
Engineering lead
See which onboarding docs are rotting and who owns them. Manage knowledge debt like tech debt.
Ops / People lead
Track owners and review dates across 50+ policy docs. Know what's stale every Monday morning.
Product / GTM lead
Keep sales decks, one-pagers, and competitive docs current. Know when a renewal brief is out of date.
Technical founder
Ship a knowledge base your AI agents can actually read — MCP server, structured YAML, no scraping.
One binary, zero dependencies. Install via Homebrew or Cargo and start building in seconds.