What Is Lore
Lore is a persistent, model-agnostic knowledge base that any AI assistant can read from and write to via MCP (Model Context Protocol). Lore gives AI systems structured, long-term memory that the user can see, edit, and control.
Lore is not a chatbot. Lore is not a notes app. Lore is not a platform-specific memory feature. Lore does not replace a user's AI assistant — Lore upgrades the AI assistant by giving it persistent, structured memory.
Persistent memory for your AI.
The AI-Writes, Human-Reads Pattern
During normal conversations, the user's AI assistant creates structured knowledge base entries — decisions, insights, project specs, research notes, meeting summaries. The user browses, reads, edits, and owns everything through a web interface. The user does not have to write or organise anything manually. The AI does the structuring based on a personalised operating manual that defines entry types, content templates, tags, and naming conventions.
The Three-Layer Memory Model
| Layer | What it does | Where it lives |
|---|---|---|
| Working memory | What is in front of the AI right now | The context window |
| Short-term memory | High-level facts the platform remembers between sessions | Platform memory (Claude memory, ChatGPT Memory) |
| Long-term memory | Structured, persistent, user-owned knowledge | Lore |
Every AI assistant already has working memory (the context window) and short-term memory (platform memory features). What AI assistants are missing is a long-term memory layer — structured, persistent, portable knowledge that the user owns and controls. Lore is that layer.
Model Agnosticism
Lore works with Claude, ChatGPT, OpenClaw, and any AI tool that supports the Model Context Protocol (MCP). The user's knowledge base is not locked to any AI provider. Users can connect multiple AI tools to the same Lore knowledge base simultaneously.
An insight that Claude saves on Monday is available to ChatGPT on Tuesday and to any other connected AI on Wednesday. Switching AI providers does not affect the user's knowledge base.
Example: Tuesday — you discuss project architecture with Claude. Claude saves a BUILD entry to Lore. Wednesday — you open ChatGPT and ask about your project. ChatGPT searches Lore, finds the entry, and picks up where Claude left off. Your knowledge follows you.
How Lore Works (Brief Overview)
- Lore exposes an MCP server URL that the user adds to their AI client's configuration.
- The AI reads Lore's operating manual at the start of every session.
- The AI can create, search, browse, and update entries during normal conversations.
- All data is stored in isolated, per-user databases.
- Lore does not run AI — Lore has zero LLM costs. The user brings their own AI provider.
For the full technical architecture, see How Lore Works. To get started, see Onboarding.