If you back founders early, let's talk.
We are not running a process. We are building. But we are listening — and if there is a fit, we want to know early. Read the thesis. Then open a conversation.
The next $10B knowledge tool will be built for AI — not retrofitted for it.
Every knowledge worker paying for Claude opens it with the same problem: their notes were not built for AI. They burn 500–2,000 tokens per session re-establishing context. A hundred sessions a month is 50K–200K tokens wasted. Multiply across 20M+ paying users and the spend is staggering — and entirely solvable with the right substrate underneath.
Obsidian was built for humans writing markdown in 2020. Notion was built for collaborative docs in 2016. Neither was designed with AI as a first-class consumer of the data. Marje is. Native MCP server, two-tier factory configuration, hybrid storage, cascading folder rules, and a curated marketplace — designed for the model before the model was a given.
We are attacking from the consumer side first. “Can my mom use this?” is the test. Enterprise governance follows the same architecture — the factory layer already enforces it. The early access waitlist is open and shaping the rollout.
Six reasons this is hard to replicate.
Two-tier factory configuration
A hidden, AES-256-GCM encrypted layer governs routing, permissions, and rendering. Users customize the surface; the system stays governed. The factory layer protects against prompt injection, conversational probing, and competitive reverse engineering.
MCP-native by design
A built-in Model Context Protocol server with 50+ purpose-built tools. Designed for Claude from the first line of code — not retrofitted with a plugin. First-mover in AI-native local-first knowledge management.
The token-efficiency moat
Architectural pre-loading eliminates the 50K–200K monthly tokens a Claude+Obsidian user burns re-establishing context. The vault pays for itself in saved API spend. Hard to replicate without rebuilding the foundation.
Patent Pending IP
Provisional patent filed on the two-tier factory/user configuration architecture and the cascading rule engine. Trademark in progress. AES-256-GCM factory rules plus native binary distribution make piracy and reverse engineering prohibitively expensive.
The marketplace
Folders that act like apps — Client List, Job Estimates, Podcast Tracker. Three categories: Workspaces, Intelligence, Data Forms. Curated, signed packages distributed via Cloudflare R2 with per-user entitlement tokens. Two-sided market on deck.
Zero variable AI cost
Customers bring their own Claude, OpenAI, or local Ollama. Marje captures 100% of license and module revenue with no API cost-of-goods. Launch infrastructure runs at effectively $0 on Cloudflare free tier.
Zero variable cost. Multiple revenue lanes.
Where we stand today.
- Product — Shipping on Windows. Daily build, single founder, full vertical product (Svelte frontend, Rust backend, MCP server).
- Platforms — Windows today. macOS within 60 days. Linux when the signal shows up.
- IP — Provisional patent filed. Trademark in progress.
- Stack — Tauri v2, Svelte 5, Rust. Cloudflare for licensing and module distribution (R2, Workers, D1).
- Distribution — Direct download today. Module store runs on Stripe Checkout with signed entitlement tokens.
- Infra cost — ~$0 per month on Cloudflare free tier. Code signing and distribution only.
- Enterprise — Foundation shipped (audit, RBAC, approvals). Centralized policy and SSO in early access.
- Founder — Kevin Hibbard. Owosso, Michigan. Inventor, full-stack developer, serial entrepreneur. Multiple operating businesses under Hibb Co. LLC.
Open a conversation.
Tell us about you. We are not running a process — we are listening. If there is a fit, we move fast. If there isn't, we tell you straight.
Prefer email? [email protected].