AI Computer reComputer RK3588-30 8GB LPDDR5 - Seeed Studio 100071234

Description

AI computer Seeed Studio reComputer RK3588-30 uses the Rockchip RK3588 with four Cortex-A76 cores clocked at up to 2.4 GHz and four Cortex-A55 cores running at up to 1.8 GHz. The processor works with the Arm Mali-G610 MC4 GPU, 8 GB LPDDR5 memory, and an NPU providing performance of up to 6 TOPS. The hardware accelerator enables local execution of models for image analysis, speech recognition, text generation, and other edge AI applications.

The platform supports models and tools used for CV, LLM, VLM, STT, and TTS tasks. The reComputer AI Lab environment provides access to ready-to-use models and deployment tools for Rockchip, including RKNN-Toolkit2. Supported solutions include YOLO, MobileNet, RetinaFace, CLIP, Whisper, DeepSeek-R1, and Qwen2.5-VL, while compatibility with ONNX, PyTorch, TensorFlow, and Caffe facilitates migration of existing projects to the RK3588 platform.

The built-in NPU can be expanded with an accelerator connected through an M.2 slot designed for an NVMe SSD or AI accelerator. According to the manufacturer, using an additional accelerator can increase the total AI performance up to 26 TOPS. The miniPCIe connector also enables expansion with communication modules such as 4G LTE, LoRaWAN, or Wi-Fi HaLow.

reComputer RK3588-30 provides two 2.5 Gigabit Ethernet ports and four USB 3.0 Type-A ports. Available interfaces also include USB Type-C with OTG and DisplayPort, connectors for cameras and displays, GPIO, a microSD card slot, nano SIM slot, and expansion interfaces. The extensive set of wired interfaces and expansion options enables integration with cameras, peripheral devices, data storage, and communication modules.     

The multimedia subsystem supports hardware video decoding up to 8K at 60 fps and video encoding up to 8K at 30 fps. Available video interfaces include HDMI and DisplayPort over USB Type-C, while the platform supports up to four displays simultaneously. The multimedia capabilities allow the computer to be used in image analysis systems, multi-display operator workstations, and applications processing high-resolution video streams.

The pre-installed Armbian environment allows system configuration to begin after connecting power and peripheral devices. Active cooling supports processor operation under sustained loads. reComputer RK3588-30 can be used as an edge AI computer in vision systems, robotics, automation, IP camera image analysis, local AI assistants, monitoring systems, and industrial applications requiring local data processing.

Features

• Processor and memory
  • Manufacturer: Seeed Studio
  • SKU: 100071234
  • Model: reComputer RK3588-30
  • SoC: Rockchip RK3588
  • CPU: 4 × Cortex-A76 up to 2.4 GHz + 4 × Cortex-A55 up to 1.8 GHz
  • Number of CPU cores: 8
  • GPU: Arm Mali-G610 MC4
  • RAM: 8 GB LPDDR5
• Artificial intelligence
  • NPU: up to 6 TOPS
  • AI expansion: up to 26 TOPS with a compatible accelerator
  • Toolchain: RKNN-Toolkit2
  • Supported frameworks: ONNX, PyTorch, TensorFlow, Caffe
• Communication interfaces
  • Ethernet: 2 × 2.5 Gigabit Ethernet
  • USB: 4 × USB 3.0 Type-A
  • USB Type-C: OTG and DisplayPort
  • SIM card: 1 × nano SIM
• Expansion interfaces
  • M.2: support for NVMe SSDs and AI accelerators
  • PCIe interface for M.2: PCIe 3.0 x4
  • Additional M.2: PCIe 2.1 x1
  • miniPCIe: slot for 4G LTE, LoRaWAN, or Wi-Fi HaLow modules
  • Memory card: 1 × microSD
  • GPIO: GPIO expansion header
  • Camera: MIPI CSI interfaces
• Video and multimedia
  • HDMI: 2 × HDMI 2.1
  • HDMI input: 1 × HDMI 2.0
  • DisplayPort: over USB Type-C
  • Simultaneous display support: up to 4
  • Video decoding: up to 8K / 60 fps
  • Video encoding: up to 8K / 30 fps
• Power and other parameters
  • Supply voltage: 9-19 V DC
  • Buttons: Power, Recovery, MaskROM
  • Cooling: heatsink with fan
  • Operating temperature range: 0°C to 55°C
  • Operating system: Armbian

Kit includes

• reComputer RK3588-30 - 1 pc.
• 12 V / 3 A power adapter with interchangeable US / EU / UK / AU plugs - 1 pc.
• 32 GB microSD card - 1 pc.
• User manual - 1 pc.