systeric Your company's knowledge, answered by your AI.
Today every question routes through the one person who knows. Lore removes that bottleneck: your team asks in plain English, in the AI they already use, and gets the right answer in seconds, scoped to who they are.
Works with Claude, ChatGPT, and any MCP client.
Two ways to embed
Lore lives where your people already work
However your team works, Lore embeds into it, with no portal to adopt and no second tool to learn. The same grounded, scoped engine, two ways in.
In the AI you already use
Connect Lore as a connector in Claude, ChatGPT, or any MCP client. Your team asks where they already work, and every answer is scoped to who they are.
- Claude
- ChatGPT
- Any MCP client
In your own product
Embed Lore's grounded engine directly in your app, the way we did with Glide. Your product brings its data, Lore is the brain, one assistant for your users.
- Your app's assistant
- Your data as tools
- Grounded + scoped
In the AI you already use
From connected to answered, in your own client
Lore lives behind your AI assistant over an open connector. No portal to adopt, no workflow to change: your team asks where they already work, and the answer is scoped to who they are.
- 01 Connect over MCP Add Lore as a connector in Claude, ChatGPT, or any MCP client. One setup, no new app for your team to learn.
- 02 Ask in plain English In the AI they already use, anyone asks a question the way they would ask a colleague.
- 03 Get a scoped answer The right page, the table with a runnable query, the standard or process, scoped to the boxes that person is in.
- 04 Propose and review A good answer becomes a new page. A human reviews it in the editor, and the knowledge compounds.
What Lore Does
A knowledge layer your whole company can talk to
People can't find which table answers their question, or how you do a thing, so it routes through whoever knows. Lore removes that bottleneck for delivery, management, and ops alike.
Ask in plain English
Your AI assistant is the interface, Claude, ChatGPT, or any MCP client. No new app, no dashboards to learn. Ask a question and get the table and a runnable query, the standard, or the process, with the reasoning.
Curated, not scraped
Every page is reviewed knowledge, not raw documents. Answers come from a maintained synthesis, so they are trustworthy by construction instead of confidently wrong.
Answers in seconds
Knowledge is compiled once, then a vector index turns a plain-English question into the few right pages, even when the words don't match. Fast and token-efficient.
Scoped per team
Knowledge sits in groups: a company-wide box plus one per department, project, team, client, or role. People see general knowledge plus the groups they are in, never more.
Your data stays put
Lore describes your tables and processes; it never stores or runs queries against your data. Execution and permissions stay in the tools you already use.
Gets better with use
Every answered question can become a new page, reviewed by a human in the editor. Your knowledge compounds with use instead of going stale.
More than Q&A
One knowledge layer, many surfaces.
Because Lore serves curated knowledge over an open connector, the same layer powers more than a team asking questions. Point any AI surface at it.
Answers in your team's AI
The default. Your team asks in Claude, ChatGPT, or any MCP client and gets the right answer, scoped to who they are.
A customer-facing chatbot
Point a support assistant at a scoped box. It answers from reviewed knowledge instead of guesses, and deflects the repetitive questions before they reach a human.
Automated flows and agents
Let agents read Lore to act: triage and route requests, draft replies, enrich tickets, and keep workflows moving on knowledge that stays current.
Where Lore Fits
One layer, answers for every role
The machinery is the same for everyone. What changes is the box a person can see, so each role gets the answers that are theirs, in the AI they already use.
Engineering & delivery
"How does our deploy pipeline work? What did we decide here, and why?" Standards, conventions, and decisions, so engineers stop re-asking and onboarding is self-serve.
Data & analytics
"Which table has revenue by channel, net of refunds?" Table grain, joins, and metric definitions, with a runnable query ready to go.
Customer support
"What's our policy on this, and where do I check this customer's status?" Scoped to the support box, front-line staff self-serve the correct answer.
Sales
"What are our terms here, and what's the latest on this account?" Positioning and pricing from knowledge, live deal status fetched on demand, so reps prep in seconds.
Operations
"What's this request's SLA, and how is resolution measured?" Process and live status together, so ops moves without waiting on a lead.
Leadership
"How is net revenue actually defined?" Trustworthy definitions and standards before a board deck, not a data-team ticket.
Case study · Glide
Glide and Lore, answering as one
Glide holds the work: requests, cycles, roadmap. Lore holds the knowledge and the engine. Put them together and a question in plain English becomes a grounded answer that fuses your live work data with your curated knowledge, scoped to who is asking and linked to every source.
- 01 Ask in plain English Inside Glide, where the work already is. No new tool, no query language, no dashboard to build.
- 02 Lore grounds and scopes One engine decides what to pull and for whom, answers only from real results, and says "I don't know" rather than guess.
- 03 Live data meets knowledge Glide's requests, cycles, and roadmap, fused with Lore's definitions and policies, in a single answer.
- 04 Answered, with the receipts Grounded in actual records, deep-linked to each one, scoped to who is asking. Nothing crosses between teams.
What shipped this week, and did anything breach our response SLA?
18 requests shipped this week. One urgent ticket breached our 4-hour first-response SLA:
One answer: a live count from Glide, the breached request, and the SLA rule from Lore, each traceable to its source.
- Grounded, never fabricated
- Scoped to who is asking
- Linked to every source
Live demo
Ask the knowledge base
This is a live Lore instance answering questions from a sample knowledge base: support policies, onboarding guides, and business metrics. Ask anything.
Try asking
Get started
Bring Lore to your team.
Stop routing every question through a human. Let your team ask their AI assistant and get the right answer in seconds, scoped to who they are.
Talk to us