All Products

Here are the most relevant options for an “AI module,” grouped by how you want to add AI to a project:

AI accelerators (add inference horsepower to an existing host)

• For Raspberry Pi 5 via M.2 HAT+: the Raspberry Pi AI Kit (Hailo-8L, M.2 accelerator + M.2 HAT+) is a straightforward way to add fast edge inference ($120.82).
• For M.2 2280 systems needing a higher-end accelerator: the DX-M1 AI Accelerator M.2 Module - 4GB LPDDR5, 25 TOPS offers 25 TOPS (INT8) in a standard M.2 2280 form factor ($341.75).
• For a USB plug-in AI coprocessor on Linux hosts: the Google Coral USB Accelerator adds a Google Edge TPU over USB ($274.30).

AI camera / vision modules (AI runs on/with the camera)

• For Raspberry Pi intelligent vision sensing: the Raspberry Pi AI Camera (Sony IMX500 intelligent vision sensor) is purpose-built for on-sensor vision inference ($124.43), and the Raspberry Pi AI Camera - 12.3MP Sony IMX500 is a similar IMX500-based option listed at $172.10.
• For a compact MCU vision kit with an NPU: the Grove Vision AI v2 Kit - with Raspberry Pi OV5647 Camera Module and Seeed Studio XIAO is based on Arm Cortex-M55 + Ethos-U55 and targets embedded vision deployments ($71.30).
• For a plug-and-play AI vision sensor: Gravity: HUSKYLENS 2 - 6 TOPS LLM MCP AI Vision Sensor (20+ Models, Deploy Custom Models) provides 6 TOPS with multiple built-in models and support for custom model deployment ($126.70).

Embedded AI modules/boards (AI + MCU/CPU in one)

• For Kendryte K210-based edge AI: the M1 AI+lOT Module K210 Deep learning is a compact K210 AI module option ($29.90).
• For an ESP32-S3-based “AI camera module” concept: the ESP32-S3 AI Camera Module (Edge Image Recognition, Night Vision, ChatGPT Voice Interaction) combines camera + mic + speaker with edge recognition features ($40.15).
• For offline gesture/face interaction sensing: the Gravity: AI Visual Gesture and Face Tracking Sensor (5 Gestures, Range 3m) is a dedicated gesture + face tracking module ($28.85).