# VanaMD

> Knowledge management built on Git for the AI era.

VanaMD is a collaborative knowledge base built on Git versioning and Markdown. It brings the discipline that Git gives code — versioning, ownership, review workflows, structured format — to all organizational knowledge. Humans write in a rich editor; AI agents read the same Markdown files natively without conversion. A specification engine measures how completely documents state what they mean, helping teams close the gap for both human and AI consumers.

## Core Features

- **Collaborative WYSIWYG editor** — non-technical users write normally; Markdown in Git underneath is natively AI-readable without conversion
- **Git-backed versioning** — true version control (branch, merge, blame, diff) on all organizational knowledge, not just version history
- **Specification depth scoring** — measures how much of a document's meaning must be inferred vs. what is explicitly stated
- **Bidirectional epistemic trust layer** — inbound directives tell agents how much to trust each document; outbound evaluation scores agent actions before they modify content
- **Knowledge graph with gap analysis** — surfaces missing connections, stale content, contradictions, and specification gaps
- **Compliance and governance** — attestation tracking, review workflows, direction-aware classification, audit trails backed by Git history
- **Auto-generated llms.txt index** — every document discoverable by AI agents without custom integration
- **Direction-aware classification** — Reference, Operational, Published, Team, Restricted, Private with enforced governance
- **Agent policy files versioned in Git** — trust constraints travel with content, auditable via git log

## Pricing

- **Community Edition** — Free. Open-source collaborative editor with Git sync, knowledge graph, health scoring, and llms.txt generation.
- **Professional Edition** — $15/seat/month. Full specification engine: depth analysis, improvement suggestions, audience-aware scoring, contradiction detection, review workflows, SSO.
- **Enterprise Edition** — $35/seat/month. Explication engine: continuous specification monitoring, bidirectional epistemic trust layer, agent policy files in Git, source ingestion, SOC 2/FedRAMP/HIPAA.

## Status

Pre-launch. Building in the open. Go backend (Project Vanir) nearing production. Self-hosted: Docker on Linux. Hosted SaaS option planned.

## Contact

- Website: https://vanamd.com
- GitHub: https://github.com/rathbunmatt/vanamd-dev
- Email: matt.rathbun@gmail.com
- Founder: Matt Rathbun

## Essays

### What Your Search Results Are Actually Telling You (March 2026)
https://vanamd.com/blog/search-results/

Your search results aren't broken. They're the most honest picture of your knowledge base you'll ever see. The search bar strips away all the compensations you've built — your folder hierarchy, your mental model of which docs are current — and shows the raw state: flat, undifferentiated, shapeless. Every result the same kind of thing. Knowledge has personas — procedures, policies, decision records, scratch spaces — but your tools don't model any of this. The flattening makes knowledge harder to find AND harder to improve. And when AI agents query your knowledge base, they see the same flat landscape without any compensating context.

### Your Wiki Is a Camera Roll (March 2026)
https://vanamd.com/blog/your-wiki-is-a-camera-roll/

The same dynamics that turned your camera roll into 14,000 unsorted photos are about to hit your knowledge base. The marginal cost of creating a document is approaching zero, and the system has no concept of the difference between a compliance policy and a meeting summary that nobody will ever read again. Manual curation won't save you — curation disconnected from creation never scales. The next generation of knowledge tools needs a glance layer: a default view where what matters is what you see, and everything else is preserved but not competing for attention.