Skip to content

genUIne

clawfree is a hands-free AI agentic orchestrator that lets anyone create and deploy OpenClaw agents using only voice.

Code Video Cerebral Valley

Core trick

Speech input goes to Opus, which streams A2UI JSON that a Flutter client renders as native components sized to the device.

Tags

voiceagents

Description

clawfree is a hands-free AI agentic orchestrator that lets anyone create and deploy OpenClaw agents using only voice. AI agents are transformative but locked behind technical barriers—programming skills, complex UIs, and infrastructure knowledge exclude those who'd benefit most. clawfree combines voice-first interaction with Opus 4.6's real-time UI streaming for a fully hands-free experience. Voice commands on Apple Watch while jogging. Glanceable interfaces on iPhone while commuting. Full orchestration on iPad or Mac at your desk. Say "Plan a 3-day foodie trip to Tokyo" and watch Opus deliver context-adaptive interfaces—itineraries, maps, flight comparisons—all voice-triggered. Opus 4.6 as genUI Architect: Beyond chat, Opus reasons about what to present and how based on device and context. Streaming structured A2UI JSON, our Flutter client renders native components in real-time—adapting layout to screen size, interaction to input method, and density to attention context.
dart-technologies/clawfree★ 31 forksDartMITCreated Feb 10, 2026 · last push Feb 16, 2026

README excerpt

# clawfree

Hands-free AI agentic orchestrator powered by Flutter genUI

Create and deploy OpenClaw agents using only voice — no coding, no typing required. clawfree combines voice-first interaction with Opus 4.6's real-time UI streaming for a fully hands-free experience meeting users where they are: on-the-go, hands-full, eyes-busy.

## Video Demo

[![clawfree: Hands-Free AI Agent Orchestrator](https://img.youtube.com/vi/7kTwZudx40Y/maxresdefault.jpg)](https://youtu.be/7kTwZudx40Y)

## How It Works

1. **Speak** -- voice commands on Apple Watch while jogging, iPhone while commuting, iPad at the coffee shop, or Mac at your desk
2. **See** -- Opus 4.6 generates context-adaptive interfaces in real-time: itineraries, interactive maps, flight and lodging comparisons — all voice-triggered
3. **Refine** -- continue speaking to adjust; Opus adapts layout complexity to screen size, interaction patterns to input method, information density to attention context
4. **Deploy** -- save and your agent is live on OpenClaw, connected to 18+ messaging channels

## Architecture

```
Voice Input (STT)
    │
Flutter genUI Client ──── A2UI protocol ──── OpenClaw Gateway
    │                                             │
Surface widgets                              Opus 4.6
    │                                        (reasoning +
Voice Output (TTS)                           UI generation)
```

- **Flutter genUI v0.9** renders dynamic UI from A2UI JSON at runtime
- **OpenClaw** provides the AI ga

Team