Serve your docs to AI agents over SSH
# Teach your agent how to use OpenLore ssh openlore.sh teach | your-agent-cli # Add documentation access to your AGENTS.md ssh openlore.sh agents >> AGENTS.md # Explore documentation ssh openlore.sh tree -L 2 /
AI agents are trained on bash. They explore code with ls, cat, grep, and find.
OpenLore gives them the same interface for your documentation — a read-only bash shell over SSH.
No MCP servers. No RAG pipelines. No copy-pasting into context windows. Just SSH.
Pipe the teach skill directly to your agent to set up documentation access:
ssh openlore.sh teach | your-agent-cli
The teach skill walks your agent through:
Clone the repo, drop your docs in assets/lore/, build a single binary with everything baked in.
Set up lore.json with per-agent path access — different agents see different docs.
Share the binary. Anyone who runs it gets an SSH server with your documentation. Agents connect and explore.
Ship your documentation as a single binary that agents can connect to. Use the GitHub Action to automatically build and release binaries whenever your docs change.
# In your CI pipeline
- uses: aakarim/openlore@v1
with:
docs-dir: ./docs
config: ./openlore.yml
Your users' agents connect via SSH — no API keys, no client libraries, no new protocols.
Just ssh -p 2222 docs.yourproduct.com "cat /docs/api.md".
# Install go install github.com/aakarim/go-openlore/cmd/openlore@latest # Serve a directory openlore ./docs # Connect ssh -p 2222 localhost # Or run commands directly ssh -p 2222 localhost "grep -r 'authentication' /docs"
We're building toward a world where agents can access project knowledge as easily as they access code.
OpenLore uses Go's embed package to bake docs into binaries — making knowledge truly portable and distributable.
Agents sharing knowledge with other agents. Products sharing documentation with any agent. Multi-project systems where information flows freely. Open distribution of lore.