Just shipped a new feature in my Synthetic Image Engine:
👉 A controllable handwriting-style generator, especially built for CJK characters.
Indie Hackers doesn't seem to support inserting images in posts yet. If you're interested, check it out on GitHub.
Project page (showcases in GitHub):
https://github.com/alrowilde/ocr-producer
Demo(Comparing with HandRight):
https://github.com/alrowilde/ocr-producer/blob/main/handWriter/README.md
Most synthetic data tools handle printed text well.
But once real documents include handwritten Chinese/Japanese/Korean fields — things get messy fast.
This module focuses specifically on that gap.
If you're building in:
OCR
Document AI
Synthetic data
Automation involving real-world forms
Curious — how are you handling handwriting in your pipeline?
Would love to connect with others working on messy document problems.