Compact Dual OLED Display with ESP32-S3 & RP2040 Variants, Open-Source Display

 TouchEye, an innovative dual-round OLED display module designed for developers, makers, and interactive projects. Launched recently on Kickstarter by Vandana Singh, TouchEye offers a compact, feature-rich platform that comes in two versatile variants — one powered by ESP32-S3 and the other by RP2040.

TouchEye is a dual circular touch display module designed to bring vibrant visuals and intuitive touch interaction to your projects. Whether you’re building a robotic face with expressive eyes, a wearable gadget, or an IoT control interface, TouchEye’s round touchscreens provide an innovative and eye-catching user experience.

TouchEye is available in two versions, tailored to fit different project requirements:

1. ESP32-S3 Powered TouchEye – Wireless Connectivity for IoT Projects

The ESP32-S3 variant is ideal for projects requiring wireless connectivity, such as IoT dashboards, remote monitoring interfaces, and smart devices.

Key Specifications:

• Microcontroller: ESP32-S3 WROOM-1 (Dual-core Xtensa LX7 @ 240 MHz)
• Wireless Connectivity: Wi-Fi 802.11 b/g/n & Bluetooth Low Energy (LE)
• Display: Dual 1.28” round touch displays (IPS LCD)
• Resolution: 240 x 240 pixels per display
• Storage: MicroSD card slot for data logging and storage
• User Input:
• Two programmable buttons
• Boot and reset buttons
• Power & Charging:
• USB Type-C interface for programming and power
• Battery connector with charging management
• GPIO Expansion: Breakout pins for connecting additional peripherals
• Extras: Power status LED
• Operating Voltage: 3.3V

2. RP2040 Powered TouchEye – Performance & Simplicity for Embedded Applications

The RP2040 variant offers high-performance dual-core processing in a low-power, cost-effective package, perfect for embedded systems, robotics, and portable devices.

Key Specifications:

• Microcontroller: RP2040 (Dual-core ARM Cortex-M0+ @ 133 MHz)
• Display: Dual 1.28” round touch displays (IPS LCD)
• Resolution: 240 x 240 pixels per display
• Storage: MicroSD card slot for data logging and storage
• User Input:
• Two programmable buttons
• Boot and reset buttons
• Power & Charging:
• USB Type-C interface for programming and power
• Battery connector with charging management
• GPIO Expansion: Breakout pins for connecting additional peripherals
• Extras: Power status LED
• Operating Voltage: 3.3V

With two circular touch displays packed into a tiny footprint, TouchEye is perfect for space-constrained projects where visual output and touch control are key. Both hardware and firmware are fully open-source, giving developers complete freedom to customize and extend the platform for unique applications. With Wi-Fi and Bluetooth LE built-in, the ESP32-S3 version allows wireless interaction with smartphones, tablets, and cloud services. 

TouchEye’s USB Type-C interface supports both power and programming, while the battery connector with charging management allows portable, standalone operation. The MicroSD card slot adds convenient storage for data logging and asset tracking.

TouchEye is currently available for backing on Kickstarter, with early bird pricing starting at just £49. For enthusiasts and professionals looking to scale their projects, combo packs, developer packs, and bulk orders are also available.

LLMStick A Compact AI USB Device Based on Raspberry Pi Zero W and LLaMA CPP

 LLMStick is a cutting-edge USB device that leverages the Raspberry Pi Zero W and LLaMA CPP to provide a portable and efficient AI solution. Designed for developers, researchers, and tech enthusiasts, this device is perfect for experimenting with large language models (LLMs) in a compact and affordable package.

Specifications:

• Hardware Platform: Raspberry Pi Zero W
• Processor: Broadcom BCM2835 ARM11, 1GHz
• Memory: 512MB RAM
• Storage: MicroSD card slot for expandable storage
• Connectivity: - USB 2.0 - Wi-Fi 802.11 b/g/n - Bluetooth 4.1
• Power Supply: 5V via USB
• Dimensions: Ultra-compact and lightweight
• Software: Optimized LLaMA CPP for AI and LLM tasks

Key Features:

• Portability: Its small size makes it easy to carry and use anywhere.
• Performance: LLaMA CPP optimization ensures efficient processing of AI tasks.
• Ease of Use: Plug-and-play functionality via USB connectivity.
• Customizability: Open-source compatibility allows for extensive customization.

Applications:

LLMStick is ideal for:

• AI Prototyping: Develop and test AI models in a portable environment.
• Educational Tools: Introduce students to AI and LLM concepts with a hands-on device.
• Research Projects: Conduct experiments with large language models on a budget.
• DIY Projects: Enhance custom tech projects with AI capabilities.

LLMStick is a game-changer in the world of portable AI devices. By combining the power of Raspberry Pi Zero W and LLaMA CPP, it offers an affordable, efficient, and customizable solution for AI enthusiasts. Whether you're a developer, educator, or hobbyist, LLMStick is a tool worth exploring.The project features a customized version of llama.cpp and setup guidelines for configuring the Pi Zero as a USB device, all available on Pham Tuan Binh’s GitHub repository.

PocketBeagle 2- BeagleBoard.org's Latest SBC with Dual-Core Processing

 BeagleBoard.org has introduced the PocketBeagle 2, a compact single-board computer (SBC) that integrates Texas Instruments' AM6232 dual-core Cortex-A53 processor and an MSPM0L1105 Arm Cortex-M0+ microcontroller. This new iteration offers enhanced performance, increased memory, and advanced debugging features, making it ideal for developers, hobbyists, and students.

Key Features:

• Processing Power: The AM6232 SoC provides a dual-core 64-bit Arm Cortex-A53 CPU at 1.0 GHz and a 400 MHz Arm Cortex-M4F real-time core, ensuring efficient performance for various applications.
• Secondary Microcontroller: The onboard MSPM0L1105 microcontroller operates at 32 MHz, offering 32KB flash memory, 4KB SRAM, and a 12-bit ADC for precise analog signal processing.
• Memory and Storage: The PocketBeagle 2 offers flexible storage solutions, with 512 MB LPDDR4 SDRAM at 3200 MHz and a microSD card slot supporting both 1.8V and 3.3V cards.
• Connectivity and Expansion: A USB Type-C port facilitates power and data transfer, while 72-pin expansion headers provide access to 8 analog inputs, 52 digital I/Os, high-speed USB, and various digital interface peripherals, enabling extensive customization.
• Debugging and Indicators: Features include a UART port compatible with Raspberry Pi Debug Probe, a 10-pin JTAG TAG-CONNECT footprint, a power button, user/boot button, and multiple LED indicators for power, battery charging, and user-defined functions.
• Power Management: Supports a 5V/1A input via USB-C, alternative power input through the Cape Header VIN, and includes a built-in LiPo battery charger for versatile power options.

Measuring 55 x 35 mm and weighing 12.7 grams, the PocketBeagle 2 maintains a compact form factor suitable for embedded applications and IoT projects. While the specifications do not list onboard eMMC flash storage, the board's design includes a footprint for a 4GB eMMC flash, indicating potential for future enhancements.

For software support, BeagleBoard.org provides Debian-based distributions tailored for the PocketBeagle 2. Users can download the latest Debian 12.9 (Minimal) image, released on February 7, 2025, from the official BeagleBoard website. As an open-source hardware platform, comprehensive documentation, including schematics and board files, is accessible on the BeagleBoard documentation page.

The PocketBeagle 2 is now available on Digikey and Mouser for $29.00. The official product page may have more information.

OpenWrt One AP-24.XY Brings Open Source Networking to a Custom Router

The OpenWrt project has introduced the OpenWrt One AP-24.XY, a custom router designed to bring open-source networking solutions to users. This new device is tailored for enthusiasts and professionals seeking a highly customizable and open-source networking platform. The router has advanced features and hardware specifications, making it a versatile choice for various networking applications.

Features

The OpenWrt One AP-24.XY is built with a focus on flexibility and performance. It supports the OpenWrt firmware, which is known for its robust customization options and extensive package repository. The router is designed to cater to both home and enterprise environments, offering features such as:

• Dual-Band Wi-Fi Support: The device supports both 2.4 GHz and 5 GHz bands, ensuring compatibility with a wide range of devices and providing optimal network performance.
• High-Performance Hardware: The router is powered by a capable processor and comes with ample RAM and storage, allowing it to handle multiple connections and advanced networking tasks efficiently.
• OpenWrt Firmware: The inclusion of OpenWrt firmware enables users to customize their network settings, install additional packages, and optimize the router for specific use cases.
• Gigabit Ethernet Ports: The device is equipped with multiple Gigabit Ethernet ports, ensuring high-speed wired connections for devices that require stable and fast internet access.
• USB Ports: The presence of USB ports allows for the connection of external storage devices or peripherals, expanding the router's functionality.

Open-Source Networking Advantages

The OpenWrt One AP-24.XY leverages the power of open-source software, providing users with unparalleled control over their network. OpenWrt's active community and extensive documentation make it easier for users to troubleshoot issues, implement new features, and stay updated with the latest advancements in networking technology.

This router is ideal for:

• Tech Enthusiasts: Individuals who enjoy tinkering with network settings and exploring advanced configurations.
• Small Businesses: Organizations looking for a cost-effective and customizable networking solution.
• Developers: Professionals who need a reliable platform for testing and deploying network applications.

The Banana Pi OpenWrt One Router is available on AliExpress for $89.00, which includes a case, three DB antennas, and a power adapter. On Amazon, the same package is priced at $104.90.

Furthermore, according to the OpenWrt page, each purchase of a new OpenWrt One contributes a $10 donation to the OpenWrt earmarked fund at SFC.

Orange Pi AIPro (8T) A High-Performance Single-Board Computer for AI and Edge Computing

The Orange Pi AIPro (8T) is a state-of-the-art single-board computer (SBC) tailored for AI and edge computing applications. Built around an advanced Huawei Ascend AI quad-core 64-bit processor, this SBC delivers up to 8 TOPS (INT8) of AI inference performance. For users requiring even greater computational power, a 20 TOPS (INT8) variant of the SoC is also available.

Designed to meet the demands of developers, researchers, and tech enthusiasts, the Orange Pi AIPro (8T) combines robust hardware with versatile connectivity options. With support for up to 16GB LPDDR4X RAM and multiple storage configurations, this SBC is a powerhouse for AI, machine learning, and IoT projects.

Specifications

Processor and AI Capabilities

SoC: Huawei Ascend quad-core 64-bit processor (likely Ascend 310B)

• Delivers up to 8 TOPS (INT8) AI performance
• Integrated unnamed 3D GPU (likely Mali-G57)
• Arm Cortex-A76 equivalent cores for efficient processing

Memory and Storage Options

System Memory: 8GB or 16GB LPDDR4X (3200 Mbps)

Storage:

• 32MB SPI flash
• MicroSD card slot for expandable storage
• eMMC flash module connector (supports up to 256GB)
• M.2 Key-M 2280 socket for NVMe or SATA SSD

Video and Audio Support

Video Output:

• 2x HDMI 2.0 ports (supports 4K resolution at 60 FPS)
• 2-lane MIPI DSI FPC connector for display connectivity

Camera Interfaces:

• 2x 2-lane MIPI CSI connectors (compatible with Raspberry Pi cameras)

Audio: 3.5mm headphone + mic jack

Networking and Connectivity

Networking:

• Gigabit Ethernet RJ45 port for high-speed wired connections
• Dual-band WiFi 5 and Bluetooth 4.2 module (with 2 antenna connectors)

USB Ports:

• 2x USB 3.0 ports
• 1x USB 3.0 Type-C port
• 1x micro USB port (for UART console)

Expansion and Additional Features

Expansion: 40-pin GPIO header for custom projects

Miscellaneous:

• Power and Reset buttons
• 2x “startup mode toggle key” (for boot device selection)
• 12V fan connector for thermal management

Power Supply:

• Up to 20V (65W) via USB-C PD port
• 2-pin header for LiPo battery (on-board charging IC)

Physical Dimensions

• Dimensions: 107 x 68 mm
• Weight: 82 grams

Initially launched in China, the Orange Pi AIPro (8T) is now available worldwide. The product page has been translated into English, and the user manual, along with links to Ubuntu and openEuler operating system images, is accessible in the Download section.

You can purchase the Orange Pi AIPro (8T) with 16GB RAM on AliExpress for $143.99. It will also be available on Amazon shortly.

For those seeking even higher performance, Orange Pi has introduced the Orange Pi AIPro (20T), featuring a 20 TOPS Huawei Ascend SoC and up to 24GB RAM. While the product page is live, this model is not yet available for purchase on AliExpress or Amazon.

The Orange Pi AIPro (8T) is a standout choice for AI and edge computing enthusiasts due to its powerful hardware, flexible storage options, and extensive connectivity features. Its compact design and robust performance make it ideal for a wide range of applications, including:

• AI and machine learning projects
• Computer vision and image processing
• IoT and smart device development
• Robotics and automation

The Orange Pi AIPro (8T) is a game-changing single-board computer that brings exceptional AI performance and versatility to the table. With its global availability, comprehensive documentation, and support for multiple operating systems, it’s an excellent choice for developers and hobbyists alike.

Keep an eye out for the upcoming Orange Pi AIPro (20T), which promises even greater performance for advanced AI applications. Whether you’re a seasoned developer or a tech enthusiast, the Orange Pi AIPro (8T) is your gateway to the future of AI computing.

Explore the possibilities with the Orange Pi AIPro (8T) today!

Enhance Raspberry Pi 5 with Adafruit PioMatter: HUB75 RGB LED Matrix Support

The Raspberry Pi 5 features improved performance due to its upgraded CPU, GPU, and quicker I/O capabilities. Nonetheless, this progress brought about a challenge: the conventional technique for driving HUB75 RGB LED matrices became unsuitable due to the transition in GPIO control from the Broadcom processor to the RP1 peripheral controller.

Adafruit created the PioMatter library to tackle this issue, utilizing the Programmable I/O (PIO) blocks found in the RP1 chip—akin to those in the RP2040 microcontroller. This advancement enables the management of HUB75 RGB LED matrices using the Raspberry Pi 5. The GitHub platform hosts the PioMatter library.

The instructions to install the Adafruit Blinka Raspberry Pi 5 PioMatter library (or just PioMatter for short) can be found on Adafruit Learn. Once you have installed the RGB Matrix Bonnet or RGB Matrix HAT to the Raspberry Pi 5, and connected a HUB75 RGB matrix, you can install the library as follows on Raspberry Pi OS:

1 2 3 4 5 6

python -m venv ~/venvs/blinka_venv source ~/venvs/blinka_venv/bin/activate pip install adafruit-blinka pip install pillow pip install numpy pip install Adafruit-Blinka-Raspberry-Pi5-Piomatter

You’ll also need to add the PIO subsystem rule configuration by editing /etc/udev/rules.d/99-com.rules:

1	SUBSYSTEM=="*-pio", GROUP="gpio", MODE="0660"

After a reboot, you should be OK to run some commands from the terminal to drive your HUB75 RGB matrix. For example, the script below will draw a square, a circle, and a triangle:

1	python single_panel_simpletest.py

Adafruit Learn has more examples explaining how to play a GIF, run scrolling text, mirror the serial console, and even play MP4 videos like Big Buck Bunny as shown in the video below.

Please note that the PioMatter library is currently in its alpha stage, so you might encounter some bugs. If you're interested in alternative solutions, the Pimoroni Interstate 75 W (RP2040) board is designed for HUB75 LED matrix panels, though playing MP4 videos on it can be more challenging.

By integrating the PioMatter library, the Raspberry Pi 5 regains its capability to drive HUB75 RGB LED matrices, opening up new possibilities for creative projects and displays.

This article is inspired by the original coverage from CNX Software