OpenEnsemble

OpenEnsemble is a self-hosted AI assistant platform for people who want shared access on their own server, with a separate workspace for each user. Linux is the main installation target; macOS supports local development with different feature support. You can choose a single assistant or an ensemble of specialist agents using different models. It's open source under AGPL-3.0.

You can connect Ollama or LM Studio for local LLM use, or use cloud providers including Claude and OpenAI. Hosting the platform yourself doesn't make every interaction local: cloud chat providers receive the conversation context sent to them. Cortex supplies local models for internal reasoning and retrieval, while the assistant uses a separate chat provider.

Saved chats and project spaces keep ongoing work together. Agents can use custom skills and external tools through MCP connections, while scheduled tasks cover reminders, recurring work, and monitoring. The Run Inspector shows background activity, and resumable jobs let you recover interrupted agent work.

Email, Google Calendar, and Home Assistant connections bring personal services into the assistant. Dashboards can combine home controls, calendar and email views, remote node status, and custom skill widgets, with layouts for tablet displays. Routines store action sequences you want to reuse.

OpenEnsemble also supports open-hardware ESP32 voice devices. Voice data follows your provider choices: remote speech recognition receives recordings, cloud chat receives transcripts and context, and remote speech synthesis receives reply text. Local speech recognition, chat, and speech synthesis keep voice processing local.