Guides
MCP
Code search and memory for every agent
Agents waste context re-reading a repository and forget project facts between sessions. One MCP server indexes your code and docs and keeps memory that follows you across clients.
$
curl -fsSL https://www.openllm.sh/install | bash$
openllm claudeWhat it gives your agent
Tools your agent can call while it works.
- Semantic search over indexed repositories, updated incrementally
- Indexed documentation sites, searched by meaning
- Memories saved in one session and recalled in the next — in any client
Manage it in one place
Everything indexed or remembered appears on your Context page, where you can prune it.
Questions
Something else? Contact us