LR—001First product
Layerum
A knowledge platform where AI answers respect real access boundaries.
The problem
Most AI knowledge tools flatten a company into one pile of documents. Layerum keeps the layers.
- Status
- Live
- Category
- Hierarchy-native RAG
- For
- Engineering teams
- Access
- layerum.com ↗
§ 01Approach
Three ideas.
- 01
Hierarchy
Knowledge is organized as a tree of entities that mirrors how a company is actually structured.
- 02
Scope
Every question is answered inside an explicit scope, with parent layers included only when asked for.
- 03
Providers
Teams bring their own model keys and choose generation and embedding models per workspace.
§ 02 — Interfaces
One account, every surface.
- REST APIScoped API keys
- SDKTyped client
- CLIBrowser sign-in, no keys
- MCP12 tools for AI assistants
Entity tree
- Company12
- Engineering48
- Platform31
- Runbooks17
- Mobile22
- Sales36
- Legal9
Terminal
$ npm i -g @layerum-team/cli $ layerum login ✓ signed in via browser $ layerum query "how do we rotate keys?" → scope Engineering / Platform
§ 03Security
Built for governed work.
- S1
Provider keys encrypted at rest
- S2
Workspace isolation
- S3
Role-based access & invites
- S4
Audit log of critical actions