FontCrafter: Turn your handwriting into a real font

Original Article ↗ Hacker News Discussion ↗

A new in-browser tool promises to turn your handwriting into an installable font, processing everything locally without accounts or server-side uploads. Commenters are intrigued by the privacy-friendly approach and potential uses—from personal stationery to preserving relatives’ writing—but report mixed results, with issues around template scanning, lack of cursive support, and inconsistent glyph alignment. The project also revives interest in handwriting itself, highlighting cultural and generational differences in cursive use and offering a welcome alternative to paid, centralized font-creation services.

Overall reception

  • Many find the idea nostalgic and fun: turning one’s handwriting (or others’, like historical samples or kids’ handwriting) into a font is seen as both playful and sentimental.
  • Several note they’ve used similar tools in the past and like having fonts that preserve family members’ writing or their own from years ago.
  • Some say their handwriting is too ugly or illegible, joking it would act as “encryption” or that the world doesn’t need to see it.

Privacy & implementation

  • The “no account, no server, 100% in-browser” design is widely praised as rare and positive.
  • A few skeptics suggest testing offline (disconnecting from the network) to verify everything runs locally.
  • The tool uses opentype.js in-browser to generate fonts.

Cursive and handwriting culture

  • Major limitation: unclear or absent support for cursive / connected scripts; several people primarily write cursive and feel excluded.
  • Debate emerges over where cursive is still taught or used (US, UK, France, Germany, Russia, etc.) and whether the issue is cultural vs. generational.
  • Some describe very stylized, variable personal handwriting that would require multiple fonts and randomization to feel authentic.

Accuracy, UX, and bugs

  • Experiences vary sharply:
    • Some report it “just works” and feels impressive, especially when scanning with phone apps and simple cleanup.
    • Others encounter serious issues: misdetected alignment, corner markers read as glyphs, letters shifted vertically, broken strokes after thresholding.
  • The template has configurable rows for upper/lowercase, but some find the constraints (only certain row combinations) limiting.
  • Pen thickness and scan resolution matter; thinner pens and certain DPIs produce broken or misaligned glyphs.
  • Several suggest better registration (more distinct marks, manual mark selection) and pre-processing (dilate filters) to improve detection.

Market context & alternatives

  • Commenters recall earlier web tools that were bought and folded into a single commercial service with subscription limits; this project is welcomed as independent, non-server-based competition.
  • Some point out related approaches: drawing directly on mobile, encoding handwriting as JS paths, or using OCR/LLM pipelines to synthesize fonts.