Water level report ·
Recording, storing and project ideas work day to day
86 %
Photo recognition, containers with QR labels and the project assistant are running; HEIC photos from the iPhone work too now.
Done
- Recognise components by photo, with a second pass that reads the markings and identifies the board by web search
- Containers with short codes and printable QR labels
- Project assistant with web search, a fixed match against the stock and a shopping list
- Installation by script in a Proxmox container, automatic updates only when tests are green, with a fallback to the previous version
- The language model shares the Mac with other applications through a queue
- Accept HEIC photos from the iPhone and convert them to JPEG
Planned
- Localisation: the interface in more languages
In a few days Ohmventory has grown from a skeleton into a tool that works day to day. For photo recognition it now uses a larger model that reads small markings on breakout boards more reliably; the project assistant stays with the small, fast model.
If you want to try Ohmventory yourself, there is a guide for the Mac.