Echo-OS is a voice-first autonomous OS that executes real-world tasks through continuous autonomous agents.
A real-time execution layer that runs on natural language input, built as the winner of an ElevenLabs, FeatherlessAI, and AoPS-backed hackathon.
- Voice → execution pipeline (not just transcription)
- Continuous agent loop (not request/response)
- Real-time reasoning + execution streaming
- Native OS control via Electron layer
- Tool use across local system and web
- Persistent memory across sessions
- End-to-end speech interface (STT + TTS)
| Layer | Purpose | Stack |
|---|---|---|
| Interface | Real-time transcripts + execution state | React + WebSockets |
| Agent | Planning, reasoning, orchestration | Python |
| Execution | Local + web task execution | Custom tool layer |
| Voice | Input + output speech pipeline | Whisper + ElevenLabs |
| Desktop | System-level control | Electron |
echo-os/
├── frontend/ # realtime UI + streaming state
├── backend/ # agent loop + orchestration
├── desktop/ # electron system layer
├── agents/ # planning + execution logic
├── tools/ # local + web integrations
└── memory/ # session + persistent stategit clone https://github.com/your-username/echo-os.git
cd echo-oscd backend
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
uvicorn main:app --reload
cd ../frontend
npm install
npm run devcd ../desktop
npm install
npm run electron