Getting started
See how Lore connects to the tools your company already uses, passes everything through the consent gate, and builds your first memory layer of decisions, processes, and tribal knowledge.
Help
How can we help?
Find answers about connecting tools, using Lore with agents, requesting access, and getting support.
Quick Help
Start with the part you are setting up.
Connecting tools
Understand how Slack, Notion, GitHub, Gmail, Linear, and Google Drive feed Lore, including what each connector reads, how consent scopes work, and how to pick which sources are ingested.
Agent access
Learn how agents query company memory through the API and MCP, how scoped install tokens limit what each tool can see, and how to wire Lore into Claude, Cursor, and VS Code.
Support
Stuck on a connection, a sync, or repo access? Reach the Lore team directly by email or book a walkthrough and we'll help you get set up.
FAQ
Common questions
Deeper answers for teams evaluating Lore or preparing an early integration: what it stores, how access works, and how agents connect.
What is Lore?
Lore is a company memory layer for AI agents. It connects to the internal tools where your team already works, extracts institutional knowledge such as decisions, processes, conventions, ownership, and the risks people learned the hard way, then makes that structured context available to agents through the API and MCP, so they stop starting every task like a brand new hire.
What tools can Lore connect to?
Lore connects to Slack, Notion, GitHub, Gmail, Linear, Google Drive, and other systems where company knowledge lives. Each connector is read-only and runs through a consent gate: Lore learns from what each connection can already see — the channels and repositories a human gave it access to — while excluded conversations, externally shared content, and anything unclassifiable are blocked before storage.
How does Lore decide what to store?
Every message passes a consent gate first. Lore reads only what your workspace has given it access to: public channels, private channels it was invited into, and conversations it is a party to. Externally shared and unclassifiable content is always blocked, admins can exclude any conversation or narrow the allowed scopes at any time, and widening access is always a deliberate human act on the platform itself. Before anything is written, secrets and API keys are stripped, author handles are pseudonymized, and an optional sensitivity filter can drop personal content. Only the approved, scrubbed text, plus the nodes extracted from it and the embeddings used to retrieve them, is kept.
Is Lore a chatbot?
No. Lore can power chat interfaces, but the core product is memory infrastructure. Its job is to hold and serve company context that any agent or tool can query through the API and MCP, not to be a single assistant you talk to.
What is MCP, and how do I connect a tool?
MCP (Model Context Protocol) is a standard interface that lets AI tools connect to external context sources. Lore is an MCP server, so Claude, Cursor, and VS Code can query your workspace's approved memory natively. On the MCP page you mint a scoped, revocable token per install and paste its config block into your tool. Each token only ever sees the slice of knowledge you grant it.
How does Lore handle private GitHub repositories?
GitHub repos are opt-in per repo. An admin selects which repositories Lore ingests and marks each as company-shared or restricted. Restricted-repo knowledge is visible only to users who actually have access to that repo on GitHub. Members link their GitHub identity and Lore matches their permissions automatically. Knowledge with missing repo metadata stays hidden by default.
Who can see the knowledge Lore extracts?
Company memory is visible to all current company members; personal memory only to its owner; team memory only to current branch members. Restricted GitHub knowledge also requires repository access. MCP tokens use the same server-side checks—the frontend is never the privacy boundary.
Does Lore train shared models on company data?
Lore is designed to use a company's data to serve that company's own agents, not to train shared models across customers. For the full detail on what is stored, blocked, and who can see what, see the Privacy page. Open the Privacy page.
Can I request a demo or join the waitlist?
Yes. Use the Book a demo button to email us for a walkthrough, or the waitlist form on the website to get early access. Early-access users can sign in directly from the Log in button on the homepage.
Still need help?
Reach out and we'll help you get started.