# omi

> Open-source AI wearable and software that captures conversations and screen context.

Omi (originally started as Friend) is an open-source AI platform combining wearable hardware with desktop and mobile apps. It records audio and screen context in real time to generate transcriptions, notes, and action items, backed by an AI chat with long-term memory. Users can buy pre-built devices or build their own hardware using off-the-shelf microcontrollers.

Website: https://github.com/BasedHardware/friend

## What it is for

- Transcribe daily meetings and in-person conversations continuously
- Query an AI assistant about things you previously discussed or viewed on screen
- Extract automated summaries and action items from conversations
- Build custom AI wearable hardware using open schematics and firmware

## Features

- Continuous real-time audio transcription with speaker diarization
- Screen capture integration on desktop platforms
- Memory-augmented AI chat interface to recall past conversations
- Companion mobile apps for iOS and Android built with Flutter
- Desktop clients for macOS and Windows
- Open-source firmware supporting nRF and ESP32 hardware
- Developer SDKs and Model Context Protocol (MCP) server support

## Pricing

Open source. Free and open-source software (MIT); pre-assembled hardware or DIY parts sold separately.

## Details

- Type: GitHub repo
- Category: [AI](https://xiry.app/topics/ai.md)
- Open source: yes
- Platforms: macOS, Windows, iOS, Android, Web, Self-hosted
- For: Developers, quantified-self enthusiasts, and professionals wanting local or open-source continuous capture.
- Alternative to: Rewind AI, Limitless, Friend (wearable)
- Tags: wearables, transcription, audio, second brain, open source hardware, flutter

## Why it was saved

- [@kodjima33, 24 March 2024](https://x.com/kodjima33/status/1772011777066442819): A real open-source alternative to closed AI wearables (Rewind, Avi Schiffmann's Tab), with a build-your-own hardware+software stack.

## Related

- [Castmagic](https://xiry.app/resources/castmagic): Turn long-form audio and video recordings into written content assets. (Web app, Paid)
- [DIY FRIEND AI Pendant Hardware and Software Stack](https://xiry.app/resources/diy-friend-ai-pendant-hardware-and-software-stack): Bill of materials and architecture to build an open-source AI necklace (Tip, Open source)
- [ElevenLabs Scribe](https://xiry.app/resources/elevenlabs-scribe): Speech-to-text API providing word-level timestamps and audio event tagging. (Web app, Freemium)
- [mlx\_whisper local transcription benchmark & usage](https://xiry.app/resources/mlx-whisper-local-transcription-benchmark-and-usage): Fast local audio transcription benchmark and Python snippet for Apple Silicon Macs. (Tip, Open source)
- [NVIDIA Parakeet TDT 0.6B V2](https://xiry.app/resources/nvidia-parakeet-tdt-0-6b-v2): Open-source 600M English speech-to-text model with fast inference and timestamps (AI model, Open source)
- [Tab](https://xiry.app/resources/tab): Wearable AI companion that listens to your day and offers context-aware advice (Tip)

Updated 10 October 2026. Page: https://xiry.app/resources/omi

Make your own library: Xiry takes your saved reels, tweets, Reddit bookmarks and Watch Later videos, makes them searchable and pulls the useful stuff out, like this library. It's invite-only for now; join the waitlist at https://xiry.app/#waitlist.