OpenLore is a shared knowledge repository for teams and their agents. It serves Markdown over SSH and MCP, so agents in different repositories can read the same current context, with identity-scoped permissions and governed writes.
Set up against your repository
From your existing repository, give the setup instructions to your coding agent:
ssh openlore.sh setup | amp
This fetches instructions, not a remote installer. Your agent interviews you and inspects your repository to create a separate, customer-owned <team>-lore deployment project. It configures and verifies a local server before you deploy a shared one. You do not need to hand-write openlore.yml first. See repo setup for the workflow and the other instruction commands.
What to read next
- What is OpenLore explains when shared knowledge helps and how serving, access and writes fit together.
- Ways to use OpenLore starts with agent-driven repo setup, then covers transports and packaging. For a local trial instead of a deployment project, use the Quickstart.
- Deploy a shared server covers deployment requirements and platform guides after local setup.
- Docsets and Auth explain which knowledge each identity can see.
- Writing and publishing and Folder rules explain how agents contribute safely.
- Use Commands & CLI, the configuration reference, and the rules standard library for lookup, not as prerequisites for setup.