Skip to main content

Getting Started Guide

This guide helps you install OpenHuman, complete the in-app onboarding, and run your first request.

OpenHuman is open source under the GNU GPL3 license. The codebase is at github.com/tinyhumansai/openhuman.

System Requirements​

OpenHuman runs on macOS, Windows, and Linux desktops. 4GB+ RAM recommended; 16GB+ if ingesting very large mailboxes or repositories, or running local models on the same machine.

Permissions​

When first launching OpenHuman, the OS will prompt for required permissions (Accessibility on macOS, Input Monitoring for voice hotkeys, and Camera/Microphone if you plan to use the meeting assistant). You can view and adjust these permissions anytime under Settings → Automation & Channels.

1. Download & Install​

Get the OpenHuman desktop app from https://tinyhumans.ai/openhuman or via your platform's package manager. After installation, open the app.

2. Sign In​

The first screen is "Sign in! Let's Cook". Multiple sign-in options are provided, including social login. There's also an Advanced panel if using a custom core RPC URL; most users can ignore it.

No permanent lock-in. Signing in does not grant OpenHuman ongoing access to any content. All third-party access requires explicit OAuth approval in the steps that follow.

3. Run Your First Request​

After Gmail is ingested (the first auto-fetch tick happens within twenty minutes), try the following prompts:

Briefing

  • "What do I need to know from the last 12 hours?"
  • "What's waiting for me?"

Cross-source queries

  • "Summarize what I missed today"
  • "What were the key decisions this week?"
  • "Extract action items from recent conversations"
  • "What did Sarah say about this project in email and chat?"

OpenHuman automatically selects the right model for each task.

4. Open the Obsidian Vault​

The Memory tab has a View vault in Obsidian button. Clicking it opens {workspace}/wiki/ in Obsidian. You can browse agent summaries, add your own notes, or even build manual links — the agent will pick up your edits on the next ingestion.

5. Make the Mascot Do More​

Now that the agent has memory and a model, the rest of the product is about giving it more interfaces:

  • Meeting Assistant - Drop in a Google Meet link, and the mascot joins as a real participant: listening, taking notes into the memory tree, speaking in the call, and using tools in real time during the meeting
  • Auto-fetch - Connect more sources from Settings; the scheduler pulls new data into your tree every twenty minutes
  • Native Voice - Press-to-speak input and TTS replies, so you can talk to OpenHuman instead of type
  • Subconscious Loop - Let the mascot continue working on ongoing tasks while you're away

Join the Community​

OpenHuman is in early beta. Feedback and contributions are especially valuable at this stage.

Next Steps​