A utilitarian screen translation tool for Japanese games
Current translation solutions for Japanese games are too messy for general use. VETS hopes to be a fully-featured, unintrusive, utilitarian solution for tackling that text-heavy imported JRPG or VN with confidence.
VETS uses Google Cloud Vision to spot text, for romaji conversion, and DeepL machine translation.
- Unintrusive: No overlays, drag-to-capture shenanigans or multi-window setups. Fix VETS to an area of your screen, and update translation with a keypress. Translation is organized into an ordered list of 'packets', one for each block of text.
- Universal: VETS works through screen capture: if it's visible on your screen, then VETS can read it for you.
- Romaji Processing: Optional phonetic transcription, for those learning to read!
- Cutting-Edge OCR and Translation: The gcloud and DeepL set up might be a little involved, but the end result really is the best of the best screen reading.
VETS uses cloud-based services to function, and you'll have to set them up on your side due to API access limits :(
This should be enough of a guide to do so, and fill out the 'CREDENTIALS' field in the VETS settings menu.
On Linux, you need libxcb
, libxrandr
, and dbus
in order for VETS to capture your screen. Odds are you already have them installed.
Debian/Ubuntu:
apt-get install libxcb1 libxrandr2 libdbus-1-3
Alpine:
apk add libxcb libxrandr dbus
ArchLinux:
pacman -S libxcb libxrandr dbus
The requirement is as follows: to have a Google Cloud Project on your Google account, and to have the Vision AI service activated on that project.
https://cloud.google.com/vision/docs/setup
You can follow the guide above up to installing and running the gcloud CLI tool:
gcloud init
Once that's done, you're good to go! Just run the commands shown in the VETS settings menu in order to get your credentials.
Note: The Google Cloud Token gives temporary access: it's refreshed every hour or so. Make sure to run
gcloud auth print-access-token
again if you get a warning from VETS that you need to check your credentials!
This is much simpler: you just need to sign up to the DeepL API through the link below. It should immediately display your personal token as soon as you're signed up.
https://www.deepl.com/pro-api?cta=header-pro-api
All done! Please feel free to leave an issue or reach out on Discord if you're having trouble with setup.