Skip to content

Releases: laurenvil/Uno-QClaw

QClaw v1.0.2 — App Lab Edition

09 Jun 03:43
@Draider2001 Draider2001
v1.0.2
6121445

Choose a tag to compare

View all tags
QClaw v1.0.2 — App Lab Edition Latest
Latest

QClaw v1.0.2 — App Lab Edition for Arduino Uno Q

On-device agentic AI for the Arduino Uno Q, distributed as a one-click Arduino App Lab application. The agent — local llama-server, 8 tools, 15-skill workspace — runs entirely on the QRB2210, with the chat surface served as an arduino:web_ui brick at http://<uno-q-ip>:7000.

Highlights

  • One-click App Lab install. First launch downloads the Qwen3.5-0.8B-Q4_0 weights (~537 MB) and brings up the local llama-server on 127.0.0.1:8083.
  • Two-engine model routing. Switch between on-device yzma inference and Claude API models (Opus 4.7 / Sonnet 4.6 / Haiku 4.5) from the WebUI settings panel. The agent loop, tools, and 15-skill workspace are identical regardless of which model is active.
  • Editable agent tree. SOUL.md, IDENTITY.md, AGENTS.md, USER.md, memory/, skills/, sketches/, python/ are all exposed under agent/ via the host bind-mount so the App Lab UI can edit any of them.
  • Eight built-in tools. arduino (sketch compile + flash), camera, sysfs_led, network, i2cdetect, plus workspace-scoped read_file / write_file / list_dir.
  • Optional host arduino-cli daemon. Install the systemd user unit on the Uno Q to enable the agentic compile-and-upload flow. Chat works without it.

Quick Start

App Lab:

  • 1: Download Zip from https://github.com/laurenvil/Uno-QClaw/releases
  • 2: Open the Arduino App Lab
  • 3: Go to My Apps
  • 4: Click Create New App
  • 4: Import App
  • 5: Click the QClaw App Card
  • 6: Click Run
  • 7: Then open http://<uno-q-ip>:7000 and start chatting.

Terminal:

# 1. Download
curl -L -o QClaw.zip https://github.com/laurenvil/Uno-QClaw/releases/download/v1.0.2/QClaw.zip

# 2. Import into Arduino App Lab
arduino-app-cli app import QClaw.zip

# 3. Start
arduino-app-cli app start ~/ArduinoApps/qclaw

Then open http://<uno-q-ip>:7000 and start chatting.

What's in the box
Board Arduino Uno Q (Qualcomm QRB2210, 4× Cortex-A53, 4 GB LPDDR4X)
Engine yzma llama-server (llama.cpp build 9127, ARMv8 CPU)
Default model Qwen3.5-0.8B Q4_0 (downloaded on first launch)
Cloud option Claude API via Anthropic, configurable through the WebUI
Card icon 🦞

See the full README for install steps, WebUI configuration, yzma tuning, Claude API setup, the agent/ layout, and the config schema.

Assets 3
Loading