OpenHuman

An open-source personal AI app with local model support, app integrations and multi-agent workflows, accessible through desktop, browser and terminal clients.

Alternative to:

OpenHuman is a personal AI assistant for people who want agents to act across their work apps and carry out repeatable tasks. Its orchestration engine manages multiple agents and workflows in one place.

App connections: Give agents access to Gmail, Notion, Slack, Telegram, GitHub and other supported services through integrations, MCP and Skills.
Workflow automation: Describe a task and review the agent's proposed graph on a canvas, with support for multiple agents, conditions, loops and approvals.
Schedules and events: Saved workflows can start manually, on a schedule or from an app event, and resume after a pause.
Local and cloud models: Choose Ollama, LM Studio, MLX or a local OpenAI-compatible server, or use cloud providers such as OpenAI and Anthropic.
Connected memory: Use your own CortexDB or TinyHumans' hosted CortexDB to recall documents, conversations and shared learnings and answer questions with citations; memory is off when neither engine is connected.

OpenHuman is open source under GPL-3.0, with desktop, browser and terminal interfaces. Developers can embed its shared Rust core in their own applications, and multiple agents can run in one process.

Local model backends run on your machine. Cloud model providers process requests remotely, and connections to work apps communicate with those services. TinyHumans' hosted memory requires an account and stores memory with that service.