Description
LobeHub is an open-source project on GitHub (lobehub/lobehub) that its maintainers describe as a "Chief Agent Operator," meant to organize a user's AI agents into continuous, 24/7 operation rather than one-off, single-session chats. The repository has over 82,000 stars, more than 15,900 forks, and was still receiving commits as of this writing.
Read More
The project treats agents as the basic unit of work. Its "Operator" layer hires, schedules, and reports on a user's agents, including an IM gateway so agents can be reached from chat apps the user already uses. Its "Create" layer includes an Agent Builder: a user describes what they need and the system applies an auto-configuration, with access to any supported model or modality and a library the project says has more than 10,000 skills and MCP-compatible plugins.
A "Collaborate" layer adds Agent Groups, where the system assembles multiple agents to work a task in parallel, along with shared Pages for multi-agent writing, scheduled runs, and project- and workspace-level organization for teams. An "Evolve" layer adds what the project calls Personal Memory: agents that learn from how a user works over time, built on what LobeHub describes as "white-box," structured, and editable memory rather than an opaque store.
LobeHub can be self-hosted on Vercel, Zeabur, Sealos, or Alibaba Cloud, or run from a Docker image, and supports binding a custom domain; self-hosting requires the user's own OpenAI API key. The project is written in TypeScript and is released under a custom license rather than a standard OSI license (labeled "Other" on GitHub), so anyone reusing or redistributing the code should check its terms directly rather than assume permissive open-source use. The maintainers describe themselves as a group of design-engineers building the project through a bootstrapping approach, aiming for an open, transparent product ecosystem for both everyday users and professional developers. LobeHub also maintains a Discord community, publishes a public star-history chart tracking its GitHub growth over time, and has launched twice on Product Hunt.
- CategoryAI Chatbots
- Pricing ModelOpen source
- DeploymentSelf-hosted
- PlatformsWeb
LobeHub's Core Features
Organizes multiple AI agents into ongoing operations instead of single-session chats
Includes an Agent Builder that auto-configures a new agent from a plain-language description
Provides a library of more than 10,000 skills and MCP-compatible plugins
Supports Agent Groups that assemble several agents to work a task in parallel
Offers shared Pages, Projects, and Workspaces for team-level collaboration with agents
Includes a structured, user-editable "white-box" memory layer for agents
Reaches agents through an IM gateway inside chat apps the user already uses
Self-hosts on Vercel, Zeabur, Sealos, Alibaba Cloud, or Docker
Getting Started with LobeHub
Deploy LobeHub: Self-host via Vercel, Docker, or a supported cloud, using your own OpenAI API key.
Build an agent: Describe what you need in the Agent Builder and let it auto-configure.
Connect skills: Attach relevant skills or MCP-compatible plugins from the library.
Set up an IM gateway: Reach the agent from a chat app you already use, if needed.
Group agents for bigger tasks: Use Agent Groups or Pages when a task needs several agents.
Review agent memory: Check and edit what an agent has learned through its memory layer.
LobeHub's Use Cases
- Multi-agent task automation
- Custom agent building
- Self-hosted agent operations
- Team agent collaboration
