Making AI Assistants
Cheap Again

Deploy your own AI assistant on hardware that costs just one dollar. Seamlessly bridging the digital and physical worlds.

Get Started Flash Firmware Star on GitHub ★ --
5
Platforms
9
Features Done
$1
Hardware Cost
3
OS Supported
Why RT-Claw

Intelligence at the Edge

No cloud dependency. Pure C on FreeRTOS, RT-Thread, and Linux, with LLM superpowers.

01
⚡

Pure C, Zero Bloat

No Node.js. C99 on FreeRTOS, RT-Thread, and Linux. Runs on MCUs with 400 KB RAM or any Linux host. Hardware costs one dollar.

02
🪧

Swarm Intelligence

Networked nodes discover each other, share tasks, and collaborate in real time. Each Claw is a neuron in a distributed brain.

03
🔧

LLM Tool Use

The AI dynamically orchestrates GPIO, sensors, skills, and more via function calling. No re-flashing needed.

Capabilities

Core Features

Everything you need to build intelligent embedded agents.

🤖

LLM Chat Engine

Interactive conversation with Claude API over HTTP. Chat-first shell where direct input goes to AI.

🔧

Tool Use

LLM-driven hardware control via function calling. GPIO, system info, and more dynamically orchestrated.

🧩

Skill System

Composable skills built from tools. Built-in draw skill drives LCD graphics; new skills are easy to add.

🛠

OSAL Abstraction

Write once, run on FreeRTOS, RT-Thread, and Linux. Clean claw_os.h hides all OS specifics.

🪧

Swarm Intelligence

Node discovery, heartbeat, and distributed task scheduling. Collaborative real-time execution.

🔌

Gateway Router

Thread-safe message routing between services. Decoupled, modular architecture.

🧠

Memory System

Short-term RAM ring buffer + long-term NVS Flash. Nodes learn and recall operation patterns.

⏰

Scheduled Tasks

Timer-driven execution and periodic automation. Recurring operations run autonomously.

💬

IM Integrations

Connect Feishu, DingTalk, QQ, Telegram as message channels. Control hardware from any chat.

Design

Architecture

Clean layered design with OSAL bridging multiple OS platforms.

+----------------------------------------------------------------+ | rt-claw Application | | gateway | net | swarm | ai_engine | shell | sched | im | +----------------------------------------------------------------+ | skills (AI Skills) | | (one skill composes multiple tools) | +----------------------------------------------------------------+ | tools (Tool Use) | | gpio | system | lcd | audio | http | scheduler | memory | +----------------------------------------------------------------+ | drivers (Hardware BSP) | | WiFi | ES8311 | SSD1306 | serial | LCD framebuffer | +----------------------------------------------------------------+ | osal/claw_os.h (OSAL API) | +----------------+----------------------+--------------+---------+ | FreeRTOS (IDF) | FreeRTOS(standalone) | RT-Thread | Linux | +----------------+----------------------+--------------+---------+ | ESP32-C3 / S3 | QEMU Zynq-A9 (GEM) | vexpress-a9 | Native | +----------------+----------------------+--------------+---------+
In Action

See It Work

Draw skill in action — the AI composes tools to paint on LCD in real time.

RT-Claw Demo — Draw skill painting on LCD via AI
Hardware

Supported Platforms

Run on real hardware, emulate with QEMU, or build natively on Linux. Multiple OS backends via a single OSAL API.

ESP32-S3

Xtensa LX7 FreeRTOS WiFi PSRAM BLE
  • Dual-core 240 MHz, 16 MB Flash + 8 MB PSRAM
  • WiFi + PSRAM for real-world deployment
  • Web Flash supported from this page
  • QEMU + open_eth networking available

ESP32-C3

RISC-V FreeRTOS ESP-IDF WiFi BLE
  • 160 MHz single-core, 400 KB SRAM
  • Lowest cost option (~$1 modules)
  • AI + Ethernet verified on QEMU
  • Flash to real hardware or emulate with open_eth

QEMU vexpress-a9

ARM Cortex-A9 RT-Thread SCons Ethernet
  • RT-Thread v5.3.0 BSP
  • No hardware needed
  • Boot + Ethernet verified on QEMU
  • lan9118 user-mode networking

QEMU Zynq-A9

ARM Cortex-A9 FreeRTOS FreeRTOS+TCP Cadence GEM
  • Standalone FreeRTOS (non-ESP-IDF)
  • Full Meson firmware build
  • DHCP + HTTP verified on QEMU
  • AI conversation via api-proxy

Linux Native

x86_64 / aarch64 Linux pthreads libcurl
  • No cross-compile, no QEMU needed
  • OSAL via pthreads + libcurl + file KV
  • Full AI conversation on any Linux host
  • Fastest development iteration cycle

Feature Status

FeatureDescriptionStatus
LLM Chat EngineInteractive conversation with Claude API over HTTPDone
Tool UseLLM-driven hardware control via function callingDone
Skill SystemComposable skills from tools; built-in draw skill for LCDDone
Chat-first ShellUART REPL with AI integration and UTF-8 supportDone
OSALWrite once, run on FreeRTOS, RT-Thread and LinuxDone
GatewayThread-safe message routing between servicesDone
NetworkingEthernet + HTTP on QEMU; WiFi on real hardwareDone
Conversation MemoryRAM ring buffer + NVS Flash persistent storageDone
Scheduled TasksTimer-driven task execution and periodic automationDone
Swarm IntelligenceNode discovery, heartbeat, distributed schedulingIn Progress
Skill MemoryLearn and recall frequently used operation patternsIn Progress
IM IntegrationsFeishu, DingTalk, QQ, Telegram as message channelsIn Progress
Multi-Model APIClaude, GPT, Gemini, DeepSeek, and morePlanned
Web Config PortalBuilt-in web page for runtime configurationPlanned
Claw Skill ProviderServe as a skill for other ClawsPlanned
Web Flash

Flash & Monitor

Flash firmware and open a serial terminal — directly from your browser. Requires Chrome or Edge.

Web Serial API is not available.
Please use Chrome or Edge on desktop, and make sure you're on HTTPS or localhost.

Firmware Flash

Ready
Waiting for flash...
Serial Monitor
Click "Connect" to open a serial port...
Open Source

Join the Community

GTOC (Gevico Open-Source Community) — build the future of embedded AI together.

GitHub Discord QQ Group Telegram Bilibili