Token Monitor – An ESP32-S3 desktop display that tracks AI coding assistant usage (Crowdfunding)

TokenMonitor supports Claude Code, Codex, Antigravity, and customizable dashboard views

Fractal Manifold has introduced the Token Monitor, a 4-inch ESP32-S3-based desktop display for monitoring real-time AI coding assistant usage. The device features a 4-inch 480×480 IPS touchscreen with 5-point capacitive touch, Wi-Fi 4 and Bluetooth LE 5 wireless connectivity, USB-C power, optional 1-cell Li-ion battery support via an AXP2101 PMIC, and two side buttons for power and boot functions. It shows usage information including quota consumption, session limits, token counts, reset timers, and estimated token costs. A local broker service fetches the usage data and transfers it to the device over a secure connection, keeping credentials on the user’s machine. Token Monitor specifications: SoC – Espressif Systems ESP32-S3 CPU – Dual-core Tensilica LX7 up to 240 MHz with vector extension for AI/ML workloads RAM – 512KB SRAM; 8MB PSRAM Wireless – WiFi 4 and Bluetooth LE 5 Storage – 16MB flash Display – 4-inch 480×480 capacitive touchscreen with ST7701 driver and […]

PICOTTY project enables multi-target serial remote management through Raspberry Pi Pico boards and Pi Zero 2 W SBC

PICOTTY Node

PICOTTY is an open-source project enabling serial-based remote management using a Raspberry Pi Zero 2W as a hub running a web server for the dashboard, and several Raspberry Pi Pico or Pico 2 boards with an Ethernet add-on to control and monitor the targets over serial. In some ways, it’s a multi-target IP KVM, but without video or mouse, so let’s call it an IP keyboard, since each Pico board acts as an Ethernet-connected USB keyboard, while the Pi Zero 2 W SBC manages all nodes attached to it over Ethernet. The developer, Chiranjeevi (morpheuslord), mostly uses it to control systems running Proxmox VE, but it could be useful for other headless applications where a keyboard is sufficient. Hardware for the project: Raspberry Pi Zero 2 W – Runs a web server on port 8080 and a TCP server on port 9000 to control the nodes Nodes Option 1 Raspberry […]

Espressif Systems releases ESP RainMaker Neo open-source device-to-cloud-to-phone IoT platform

ESP RainMaker Neo

Espressif Systems has just released ESP RainMaker Neo, providing a complete device-to-cloud-to-phone app stack, including the cloud backend, firmware SDK, firmware examples, phone app SDK, and reference phone app. If the name sounds familiar, it’s because it’s essentially the open-source version of ESP RainMaker cloud service introduced in 2020. Both are still built on AWS Serverless architecture, but while ESP RainMaker, now known as ESP RainMaker Classic, is managed by Espressif, ESP RainMaker Neo is hosted on the customer’s own AWS instance(s), providing more flexibility and control over the cloud architecture. While the company has made efforts to keep the same hybrid mobile app SDK and similar device-side APIs, RainMaker Neo is a new implementation with a different cloud architecture from ESP RainMaker Classic. That means existing ESP RainMaker Classic deployments cannot be directly migrated to Neo, and the open-source implementation is only recommended for new deployments. Released under the Apache […]

Antmicro releases open-source hardware Thunderbolt to dual 10GbE adapter

antmicro Thunderbolt to 10GbE adapter

Antmicro has just released an open-source hardware Thunderbolt to (dual) 10GbE adapter that serves as a reference design for a high-speed Ethernet accessory, further customizable for smart terminals, Point of Sale systems, smart cameras, IoT gateways, and more. The company has a history of releasing open-source hardware designs, and we previously covered a baseboard for Quectel SA800U-WF System-on-Module (Qualcomm Snapdragon 845) and an AMD Xilinx Kintex-7 K410T FPGA development board, both designed in KiCad. The Thunderbolt to 10GbE adapter adds to the list. Antmicro Thunderbolt to 10GbE adapter specifications: Networking – 2x 10GbE RJ45 ports via Intel X710-AT2 Ethernet controller Host interface – Thunderbolt 3 USB-C connector via Intel JHL6340SLLSQ Thunderbolt 3 controller (PCIe Gen3 x4) Misc Passive cooling via heatsink 4-pin PWM fan connector 3x LEDs Power Supply – 5V via USB-C port and DC jack Dimensions – 110 x 70 mm The company tested the adapter, I assume with iperf […]

ZecTrix Note 4 – A 4.2-inch ESP32-S3 wireless e-paper display for voice tasks

ZecTrix Note 4 4.2 inch AI e paper smart note

Designed by ZecTrix Lab in China, the Note 4 is an ESP32-S3-based AI smart note with a 4.2-inch e-paper display and built-in voice input. The device is designed to display reminders, notes, clocks, weather, images, and custom templates, while the Note 4C variant replaces the monochrome display with a four-color e-paper panel. The device features 2.4GHz Wi-Fi and Bluetooth 5.0 connectivity, a microphone, a speaker, NFC, physical buttons, a USB Type-C port, a 2,000mAh rechargeable battery, and GPIO expansion for custom hardware projects. ZecTrix Note 4 / Note 4C specifications: SoC – Espressif Systems ESP32-S3 CPU – Dual-core Tensilica LX7 up to 240 MHz with vector extension for AI/ML workloads RAM – 512KB SRAM; 8MB PSRAM Wireless – WiFi 4 and Bluetooth LE 5 Storage – 16MB flash Display Note 4 – 4.2-inch monochrome E-Ink display, 400 × 300 resolution; supports partial refresh Note 4C – 4.2-inch four-color (BWRY) E-Ink display, […]

Play games with your brain signals using Octopus 16 wireless EEG device

Octopus 16 biosignal device for gaming

Octopus 16 is a wireless biosignal HID device for the XIAO ESP32-S3 board designed to let users play computer games using brain signals (EEG) with no hands (or implants) needed. The board packs 16 electrodes to measure biosignals, including EEG (electroencephalography), EMG (electromyography), and ECG (electrocardiography), and the coin-sized device is much more compact than other brain-computer interface (BCI) solutions, which typically require a cap kit with 8 to 16 electrodes and wires. Octopus 16 specifications: ADC – 2x Texas Instruments ADS131M08 24-bit, 8-channel, simultaneous-sampling, delta-sigma (ΔΣ) analog-to-digital converters Channels – 16x 24-bit channels (8x per ADC) Programmable gain (PGA) – x1 to x128 (set in firmware; default gain register 0x2222) Input high-pass / DC block – Integrated ADC DC-block filter enabled (≈ >1 Hz) Host interface (to XIAO board) – Power + SPI via 2x 7-pin Brain interface 18x spring-loaded pogo pins 16x electrodes 1x  reference (AINREF) 1x Ground […]

reTerminal Sticky 3.97-inch touch ePaper display is supported by four open-source firmware projects (so far)

reTerminal Sticky

Seeed Studio’s reTerminal Sticky is a 3.97-inch “AI-powered” magnetic touch ePaper display powered by an ESP32-S3 microcontroller, which serves as a sticky-note alternative, an e-reader, a live dashboard, or a home automation controller. The 800 x 480 resolution display comes with 8MB PSRAM, 32MB Flash, a microSD card slot, and next/previous page buttons. It’s “AI-powered” through a microphone and an AI Voice button leveraging the ESP32-S3 Edge AI capabilities for a voice-to-note feature. It also features a temperature and humidity sensor and a 3-axis accelerometer, as well as a built-in magnet and a magnetic ring for easy mounting. reTerminal Sticky specifications: SoC – Espressif ESP32-S3R8 CPU – Dual-core Tensilica LX7 microcontroller up to 240 MHz with vector instructions for AI acceleration Memory – 8MB PSRAM Wireless – 2.4 GHz WiFi 4 and Bluetooth 5.0 LE Storage 32 MB flash MicroSD card slot (SanDisk recommended) Display 3.97-inch monochrome (B&W) ePaper 4-level […]

NightRun UEFI application boots a local LLM on Raspberry Pi 5 and x86 PCs without an OS

NightRun runs LLMs without OS

Running LLMs locally on an SBC or mini PC is a good way to keep your data private, but they typically require a conventional operating system, which uses a good portion of system memory. NightRun is an experimental open-source project that addresses this issue by booting a machine directly into a local LLM from a USB drive or microSD card, without loading a conventional OS. By removing the operating system, NightRun makes more RAM and memory bandwidth available for AI inference. The runtime is written in Rust, and during boot it loads a quantized model (1.3 GB to 2.4 GB) directly into RAM while verifying its CRC-32 checksums. After the model is loaded, storage is “sealed,” meaning any later attempt to read from the disk will trigger a hard fault. NightRun key features: Hardware support – Supports 64-bit x86_64 UEFI PCs (Secure Boot disabled) and Raspberry Pi 5 via USB […]

Exit mobile version