Self-host Lobu
lobu-v24.2.0Shared context layer for AI agents using event sourcing
- Stars
- 231
- Port
- 8787 (Web UI, API, MCP gateway)
- Requires
- PostgreSQL with pgvector extension
ghcr.io/lobu-ai/lobu-appWhat is Lobu?
Lobu is a self-hosted platform that gives AI agents a persistent, shared understanding of your organization. It captures every tool call, data change, and conversation as events tied to entities like people, projects, and customers. This creates a live knowledge graph that agents can query instead of starting fresh each session. ChatGPT, Claude, Codex, and custom agents connect via MCP to access this shared context. You can also run persistent specialist agents that handle ongoing responsibilities like customer research or support triage. Deployment options include Docker, Docker Compose, a CLI with embedded database, or Kubernetes. The project is open source under Apache-2.0.
Should you self-host Lobu?
Lobu scores 48/100 on selfhost.directory (how scoring works) and ranks #110 of 209 in AI Agents. It has 231 GitHub stars.
It shipped 10 stable releases in the past year. The latest, lobu-v24.2.0, came out on 27 October 2026. New versions land roughly every 3 days.
Setup is advanced: plan for several services and some ongoing maintenance. An official Docker image is available. It is Apache-2.0-licensed and written mostly in Typescript.
Its closest alternative on selfhost.directory is Orca (91/100, 108 815 stars). It scores higher overall, so compare the two before you pick. See Lobu vs Orca side by side.
Features
- Approval policies
- Audit Trail
- Automations
- Connectors (60+)
- Credential brokering
- Cross-source context
- Entity/knowledge graph
- Event-sourced context layer
- MCP Server
- Multi-agent delegation
- Persistent specialists
- Sandboxed code execution
- Shared organizational memory
- TypeScript SDK & CLI
Lobu README
More in Knowledge Management
LocalRecallSelf-hosted, open-source RESTful API for managing knowledge bases and vector databases, designed for AI agents and chatbots with a web UI.
LongMemoryLongMemory is an open-source, self-hosted memory engine for AI agents: temporal recall, SQLite storage, MCP server, HTTP API, CLI and a dashboard.
Many NotesSelf-hosted Markdown note-taking web app with multiple vaults, OIDC/OAuth auth, real-time collaboration, fast search, and PWA support.
marm-memorySelf-hosted persistent memory server for AI agents with 16 MCP tools across HTTP and STDIO, fusing session history, codebase indexing, and concept graphs in SQLite.
maxkbSelf-hosted, open-source AI agent platform: build RAG knowledge-base chatbots, orchestrate workflows with MCP tools, and embed Q&A into your own apps.- memExA structured personal knowledge base inspired by zettlekasten and org-mode, designed for notes, contexts, and pipelines with multi-user support and privacy controls.
MemorwiseSelf-hosted NotebookLM alternative for document chat with any LLM. Drop in files/links — local RAG Q&A, knowledge graphs, flashcards, quizzes, audio. One command.
memory-osSelf-hosted, open-source memory stack for Hermes Agent: seven layers adding Qdrant vector search, SQLite fact storage with trust scores, and automatic context injection.
Memos WorkerSelf-hosted notes and knowledge base on Cloudflare Workers with D1, KV, R2, featuring Markdown editing, full-text search, Telegram integration, and themes.
minneSelf-hosted, open-source personal knowledge management app using a graph database to auto-connect notes, with AI-assisted extraction and chat interface.
mnemoSelf-hosted AI memory sidecar with persistent knowledge graph, entity extraction, and semantic retrieval for LLM apps — works with Ollama, OpenAI, Anthropic.
moonstoneSelf-hosted headless PKM server turning Obsidian, Logseq, Zim vaults into a REST API with PWA editor, applets, and background services.
0 comments