Omi is an open-source platform for AI wearables and a comprehensive second-brain ecosystem spanning mobile, desktop, and web clients README.md5-10 The platform captures real-time conversations and screen data, runs automated speech-to-text (STT) and voice activity detection (VAD), extracts structured memories and action items via large language models (LLMs), and powers an intelligent AI chat system README.md7
The repository integrates deep firmware layers (omi/, omiGlass/) with high-level runtime services (backend/, app/, desktop/, web/) and extensible developer tools (plugins/, sdks/, mcp/) README.md132-138
For full details on system-wide architecture and component flows, see the downstream subsystem guides or review the repository structure below.
Sources: README.md5-10 README.md132-138
The repository is organized into distinct functional domains, each mapping to a primary entry point across the stack. Client applications, cloud handlers, and firmware modules each maintain dedicated entry points—such as app/lib/main.dart for the Flutter mobile client README.md133 backend/main.py for the FastAPI cloud service README.md134 desktop/macos/OmiApp.swift for the macOS desktop app README.md132 and desktop/windows/src/main/index.ts for the Windows Electron application README.md132
For a complete walk-through of directory hierarchies, bootstrap scripts (app/setup.sh, backend/README.md), and surface-specific entry points, see Repository Layout and Key Entry Points.
Sources: README.md132-165 app/setup.sh1-10 backend/README.md105-106
The documentation hub is built using Mintlify, structured around docs/docs.json docs/docs.json1-3 It categorizes content into getting started guides, core developer tutorials, hardware assembly guides, and OpenAPI-driven API references docs/docs.json11-158
Platform behavior is tightly constrained by repository invariants and failure-class governance models (defined in PRODUCT.md and AGENTS.md) README.md150 These rules dictate fail-closed safety boundaries, feature flags, and deterministic test requirements.
For a detailed breakdown of the Mintlify configuration, hardware developer paths, and product constraint rules, see Documentation Site and Product Invariants.
Sources: docs/docs.json1-3 docs/docs.json11-158 README.md150
Diagram: High-Level Ecosystem Architecture (Natural Language to Code Entity Space)
Sources: README.md103-138 backend/README.md105-106 docs/docs.json1-3
Diagram: Surface Routing and Entry Points
Sources: README.md20-40 backend/README.md105-106
Refresh this wiki
This wiki was recently refreshed. Please wait 7 days to refresh again.
