Repolode
Submit
BasedHardware

BasedHardware/omi

Open-source AI that captures your screen and conversations, transcribes them in real time, and lets you chat over what it has seen and heard.

Open-source alternative to:LimitlessOtter.ai

View on GitHubWebsite
Stars
13,774
Forks
2,636
Open issues
1,804
Last commit
3 hours ago

What is it?

Omi is an open-source personal AI that records what appears on your screen and what is said around you, transcribes the audio in real time, and produces summaries and action items. It runs across desktop, phone, and a companion wearable, and its chat feature answers questions using the captured history as memory.

The project is a monorepo with several components: a macOS app built with Swift and SwiftUI, a Windows desktop app, a Flutter mobile app for iOS and Android, a Python FastAPI backend using Firebase, firmware for the wearable (nRF/Zephyr) and an ESP32-S3 glasses prototype, device SDKs in Python, Swift, React Native and other languages, an MCP server, and a Next.js set of AI personas. The backend pipeline uses components such as voice activity detection, speaker diarization, Deepgram for speech-to-text, Firestore, Redis and LLMs.

Developers can run the macOS app against the cloud backend with a single script, or set up the full local backend stack. The repository also documents an app framework for building third-party integrations and REST endpoints for memories, conversations and action items.

Key features

  • 01Captures screen content and conversations and transcribes them in real time.
  • 02Generates summaries and action items from recorded conversations.
  • 03AI chat that answers questions using captured history as memory.
  • 04Runs on macOS, Windows, iOS, Android, and a companion wearable.
  • 05Provides device SDKs, a REST API, and an MCP server for integrations.

Use cases

  • Recalling meeting discussions and action items after a day of calls.
  • Building third-party apps or integrations on top of captured conversations.
  • Experimenting with wearable AI hardware and BLE device SDKs.

Quick start

  1. 1

    Install the requirements: macOS 14+, Xcode, and Node.js

  2. 2

    Clone the repo and build and launch the macOS app

    git clone https://github.com/BasedHardware/omi.git && cd omi/desktop/macos && ./run.sh --yolo

Things to know

  • The quick start path is for the macOS desktop app and requires macOS 14+ with Xcode.
  • By default the quick start connects to the cloud backend rather than running fully locally.
  • Running the full local backend requires configuring environment files and credentials.

Topics

#ai#transcription#wearable#flutter#speech-to-text#personal-assistant#macos#app#bci#friend

FAQ

What is omi?
Open-source AI that captures your screen and conversations, transcribes them in real time, and lets you chat over what it has seen and heard. Omi is an open-source personal AI that records what appears on your screen and what is said around you, transcribes the audio in real time, and produces summaries and action items. It runs across desktop, phone, and a companion wearable, and its chat feature answers questions using the captured history as memory.
Is omi free to use? What license does it have?
omi is open source under the MIT license. Check the license terms before using it in commercial products.
Is omi still maintained?
Actively maintained. The last commit to omi was on Oct 20, 2026.
What can I use omi for?
Recalling meeting discussions and action items after a day of calls. Building third-party apps or integrations on top of captured conversations. Experimenting with wearable AI hardware and BLE device SDKs.
How do I install omi?
Install the requirements: macOS 14+, Xcode, and Node.js Clone the repo and build and launch the macOS app: git clone https://github.com/BasedHardware/omi.git && cd omi/desktop/macos && ./run.sh --yolo
Where can I get omi?
The source code and installation instructions are on GitHub: https://github.com/BasedHardware/omi

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects