SenseCraft

AI vision

Plug-and-play AI vision models on edge devices, supports tasks like Object/Person Detection, Classification, Segmentation.

• 400+ pre-trained models, covering scenes like security, retail, traffic, etc.
• One-click deployment, no-code needed

AI sound detection

Understand and analyze sound, use words to trigger action, or identify abnormal sounds.

• One-click retrain
• Train with your computer microphone

Vibration anomaly detection

Capture subtle vibrations to predict potential failures in advance.

HMI Interface Configuration

Zero-code, easily customize smart device dashboard interfaces.

• 200+ community templates
• Connect to IoT data via SenseCraft Data
• Easy configuration and deployment for e-ink displays

Multimodal perception assistant

Perceive environment, natural dialogue, knowledge integration — your all-in-one strategy assistant.

• Embedded Chatbot and LLMs
• Works with SenseCAP Watcher

Manage a fleet of edge devices

Deploy the most popular applications to your edge devices, like OpenClaw, Home Assistant, Ollama, and LLM models.

• Bind device without installing new system
• One-click deployment

AI vibe-coding for hardware

Generate code for dev boards or SBCs, with streamlined compilation and upload experience.

Create AI applications with natural language

Describe your use case and ask AI to modify existing applications, or build from scratch.

Generate dashboard in HMI

Ask AI to generate images or dashboard designs.

Connect LLMs to your AI applications

Power your AI applications with Large Language Models provided by us, fast and affordable. No third-party platform required.

Robotics training

The best training platform for Seeed Studio's reBot robotic arms, based on LeRobot.

• Streamlined pairing, calibration, data-collecting, training, and deploying
• CPU computing power on demand

AI vision training

Use your own dataset to train AI vision or sound models, to improve reliability in specific use cases.

Convert for reComputer

Convert CV, LLM, VLM models to run directly on reComputer RK series.