After a long break, I’m back working on my FIDO2 authenticator! With BLE dropped from Chrome, I’ve shifted gears to USB—and that meant switching from ESP32 to STM32 (and soon RP2040). The latest prototype features a fingerprint scanner, secure element, and a fresh USB PCB. Still lots to build, but it’s moving again!
While working on the URU Card project, I realized just how practical the credit card form factor really is—so I redesigned URU Key into URU Card Pro. It now fits in a wallet, runs on an ultra-thin Li-Ion battery, and includes a tiny OLED display for feedback. Same secure guts, much better usability!
Hardware is done! After months of prototyping, I finally finished the compact 40×15mm URU Key device—complete with ESP32, fingerprint scanner, security chip, LiPo battery, and USB-C. It’s basically a fitness tracker… but for secure authentication. Now onto the software (and maybe a proper case 😅).
Back to FPC1020 driver work! I fired up an ESP8266 with Arduino to detect finger presence using the scanner’s built-in sensor matrix. It’s simple, fast, and could help optimize image capture—or even act like a button or gesture detector. Full example code and visuals inside!
It’s official—URU Key is now a real, standalone device! I added the custom power board, used spring-loaded pins for easy firmware updates, and screwed everything together into a compact 35×20×10mm unit. Hardware’s basically done—now it’s time to dive deeper into the firmware, especially the biometric side.
Everything’s finally coming together! I combined the ESP32, fingerprint scanner, and ATECC508A security chip into a single board—the first real version of my own FIDO2 Authenticator: URU Auth (You Are You). It’s handheld, it lights up, and it’s so close to being fully autonomous. Time to tackle power and fingerprint recognition!
Getting actual fingerprint data out of the FPC1020A scanner was way harder than I expected. After days of digging and reverse-engineering Android kernel drivers, I finally got it to scan! Now I can detect fingers and capture raw images—next stop: matching algorithms.
Finally got my hands on the FPC1020 fingerprint scanners! These things are tiny, SPI-based, and perfect for a built-in solution. I built test boards, hooked them up to a NodeMCU, and wrote a basic Arduino sketch to read the hardware ID. Next step: integrating with the main authenticator project!
Using a button for FIDO2 user verification felt… a bit too easy to spoof. So I grabbed two fingerprint modules—GROW R300 and FPC1020—to try adding real biometric authentication. R300 might even let me skip its built-in MCU and handle image processing myself. Let’s see where this goes!