Lobu

Lobu is open-source infrastructure for AI teammates that watch, remember, and act. It is built for agents that need to monitor live company data, keep a running organizational memory, and branch into a sandbox to do work based on a stated goal.

The system ingests connectors and webhooks into an append-only log, then turns those updates into live customer memory. Agents look up that memory, scan it on a schedule, and keep evidence attached to what they find. The page describes a workflow in which a user defines what to watch for and when to ask before acting; an example uses churn risk, renewal timing, and account health to trigger a drafted customer success check-in for approval. Lobu also says agents can use existing connectors or write code to connect new data sources through a Connector SDK. The listed sources include HubSpot, Stripe, Zendesk, Gmail, Google Drive, GitHub, Notion, Snowflake, PostgreSQL, and 50+ more.

It shows example uses for sales, legal, finance, leadership, market research, and community workflows. Drafted responses and agent actions can be reviewed and approved, edited, sent, or left as is. Lobu also says results can appear in a team chat or in an app over the API.

The product can be used from Claude, ChatGPT, Cursor, and OpenCode, and the site includes a setup prompt and an npx @lobu/cli@latest init my-agent command for starting a project. It supports running locally, self-hosted, or as a managed cloud service. Local use runs on a laptop, self-hosted use can run in Docker, a cloud VM, or Kubernetes, and Lobu Cloud provides managed isolation, secrets, and upgrades. The page identifies Lobu as open-source, but it does not state a separate pricing model.