You've explained yourself to your AI dozens of times.
Your role, your projects, your preferences, the way you think. Every new conversation, you're onboarding the same colleague from scratch. Not because the AI is bad at memory — because there's no structured place for that context to live and grow.
Your best conversations vanish.
That architecture decision you talked through at 11pm. The insight that reframed your whole approach. The research thread you spent an hour on. It's in a chat log somewhere — unsearchable, unstructured, decaying into noise.
You don't own your knowledge.
Whatever your AI does remember, you didn't write it, you can't read most of it, you can't structure it, and you can't take it with you. Your AI has been learning about you — but the output of that learning doesn't belong to you.
The Missing Layer
Your AI already has memory. It's missing the long-term layer.
Layer
What it does
Where it lives
Working memory
What's in front of the AI right now
The context window
Short-term memory
High-level facts the platform remembers
Platform memory
Long-term memory
Structured, persistent, owned knowledge
Lore
Your AI already has working memory and short-term memory. What it's missing is the long-term layer — a structured knowledge base that grows over time, that you can read, edit, and own.
The Shift
What if every valuable thing that came up in conversation was saved, structured, and searchable?
Not a notebook. Not a wiki. A knowledge base your AI builds for you — and you actually own. Lore isn't one thing. It's whatever you need it to be. A student's Lore looks nothing like a startup founder's, which looks nothing like a freelance artist's. The AI builds a knowledge base shaped by how you think and what you do.
How It Works
Connect. Talk. Own.
1
Connect
Paste one URL into your AI. 30 seconds. Works with Claude, OpenClaw, and anything that speaks MCP.
mcp: lore.so/mcp?token=your_token
2
Talk
Have conversations like normal. When something valuable comes up — a decision, an insight, a spec — your AI saves it to your Lore. Structured, tagged, searchable.
3
Own
Browse your knowledge base anytime. Read, edit, reorganise. It's yours — not hidden in a black box, not locked to one platform.
In Practice
Just talk. Lore handles the rest.
No new workflow. No tagging. No filing. You talk to your AI — and when something matters, it saves it.
You
I just connected Lore. What can you do with it?
Claude
I've read your operating manual — your knowledge base needs setting up. Want to go through the setup conversation? I'll learn about your work and propose entry types, tags, and a naming convention that fit.
You
We decided to go with Postgres over SQLite for the multi-tenant version. The main driver was concurrent writes.
Claude
Created: DECISION: Postgres Over SQLite for Multi-Tenant — captured the trade-offs and the concurrency reasoning.
You
Save this — we're switching the cache layer to Redis.
Claude
I found an existing entry DECISION: Caching Strategy from last week. Should I update it with the Redis decision, or create a separate entry?
You
This was a really insightful conversation. Can you summarise it in lore?
Claude
Created: INSIGHT: Why Retention Beats Acquisition for Early-Stage — captured the core argument, the data points you referenced, and the three metrics you want to track going forward.
Model Agnostic
Your lore lives in Lore. Not in your provider.
Your Lore works with Claude and OpenClaw at the same time. The insight Claude saved on Monday is available to OpenClaw on Tuesday. One knowledge base, every AI you use.
Claude
OpenClaw
+ any MCP client
persistent memory for your digital assistant
Your Lore
No two Lores look the same.
When you sign up, your AI walks you through setup. It asks about your work, proposes entry types and tags tailored to you, and builds your knowledge base. Five minutes from sign-up to a fully configured, personalised Lore.