Every AI you use forgets you — or worse, remembers you in someone else's database. LAPLAS makes your memory an editable, portable file you own; LAPLAS AX serves it to every agent in your org. The model is rented; the memory is owned.
Not notes. Not a chat log. LAPLAS captures what you think into typed nodes, properties, links, and decisions — then recalls it grounded, with the causal chain behind every answer.
recall_why walks the causal chain — derived-from, supports, supersedes — so you see why it knows, not just what..laplaspack — an open, MIT-licensed SQLite container you can read, sign, export, and leave with.
AX is the memory plane for your company's AI. Mount your team's packs, run any agent on top — grounded and cited — and let agents propose writes that a human approves into the org's canon.
Everything above runs on one open, MIT-licensed format: a SQLite container with causal provenance, offline signing, and a formal grammar. Read it with a zero-dependency script. Verify it without a server. Leave with it any time.
## Subscription [[Subscription]] >>type: offer >>status: core >>derived-from: [[Founder Interview]] $ laplaspack read --why "core offer" ◈ Subscription ← derived-from ← Founder Interview ← supports ← Pricing Principle 3 ancestors · offline · provenance intact ✓
Your memory is yours at $0, forever — pricing starts where serving it to a team does.
Deploy inside your boundary — custom harnesses & eval gates, security review, SLA. We help compile your org's memory hands-on.
.laplaspack file on your disk. Nothing leaves your machine unless you choose to sync, share a space to a workspace, or publish a Manifesto.