OpenLobster
An open-source self-hosted AI assistant with multi-user memory, secure tool use, scheduling, and multi-channel support.
🛡️ AgentReady threat assessment
MAESTRO 7-layer threat model + OWASP AIVSS risk score for OpenLobster, derived from its capabilities.
These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology. Are you the vendor? Factual corrections are free.
Overview
OpenLobster is an open-source, self-hosted personal AI assistant that evolved as an opinionated fork of OpenClaw with a stronger focus on security, multi-user support, structured memory, and operational usability. Its official GitHub repository describes a redesigned architecture with a real scheduler, per-user histories and permissions, GraphQL API, encrypted secrets storage, dashboard authentication enabled by default, and support for channels such as Telegram, Discord, Slack, WhatsApp, and Twilio SMS. It can connect to multiple AI providers including OpenAI, Anthropic, Ollama, OpenRouter, and OpenAI-compatible backends, while using Neo4j or file-based memory and offering MCP integrations with OAuth support. OpenLobster is aimed at users who want a self-hosted assistant that can operate across channels, automate recurring work, and maintain richer long-term context than simpler local agent setups.
Key features
- self-hosted
- privacy
- multi-user
- Neo4j
- GraphQL
- scheduler
- MCP
- Slack
- Telegram
- Discord
Use cases
- Running a self-hosted AI assistant across chat channels such as Slack, Telegram, Discord, and WhatsApp.
- Managing recurring and one-off automations with a real scheduler and visible task logs.
- Using structured long-term memory with Neo4j or a local file backend instead of flat note files.
- Controlling per-user permissions and secure tool access for a multi-user assistant deployment.