Skip to content

Diya-sudheer/better-text-machine

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Better Text Machine

CI License: MIT

A responsible document revision tool for source-aware, personal, and transparent writing.

Use it live — everything runs in your browser; drafts never leave your machine. At the playful entry gate, type guest.

Better Text Machine is a dependency-free writing assistant for revising rough or AI-assisted drafts into clearer, more personal, better sourced writing.

It helps writers:

  • open through a playful local "Peekaboo" name gate
  • improve clarity and flow
  • add placeholders for personal examples
  • reduce generic wording
  • suggest stronger transitions
  • flag claims that likely need citations
  • preserve the user's meaning
  • avoid plagiarism through source and disclosure checks
  • produce a revision checklist
  • import plain text or Markdown files
  • extract selectable text from uploaded PDFs in the browser
  • use free OCR fallback for scanned PDFs when Tesseract.js loads
  • import readable .docx files with Mammoth.js
  • accept writer suggestions as required revision directions
  • apply, ignore, or mark suggestions as needing rewrite
  • edit the final draft directly before export
  • track source notes beside evidence-heavy claims
  • build personal evidence from guided prompts
  • save local revision history in the browser
  • calculate a trust score from sources, personal detail, unresolved prompts, and checklist progress
  • export a true .docx file with JSZip, with .doc fallback if the library is unavailable
  • show a before/after sentence diff
  • map claims that need evidence
  • identify personalization gaps
  • score how well the revision reflects the user's suggestions
  • audit risky claims with a "do not invent facts" check
  • show a citation-needed heatmap
  • open a print view that can be saved as PDF

Open index.html in a browser to use it. PDF extraction, DOCX import, and OCR use free CDN libraries, so the browser needs internet access when loading the page.

PDF extraction works best for PDFs with selectable text. Scanned image PDFs use browser OCR, which can be slower and less accurate. Legacy .doc files are not supported, but .docx files can be imported.

This tool is intentionally not designed to bypass AI detectors. It is meant to support authentic, transparent revision and responsible final authorship.

The Peekaboo gate is a playful local entry screen, not real authentication — guest always works on the hosted demo. Change allowedGateNames in app.js if you want different names.

Security notes

  • The app uses a browser Content Security Policy to limit scripts, connections, frames, forms, and embedded objects.
  • Uploaded files are processed locally in the browser and limited to 25 MB.
  • OCR is capped to the first 6 PDF pages to avoid freezing the browser.
  • Local revision history and source notes are stored only in browser localStorage.
  • The Peekaboo gate is not real security or authentication.
  • CDN libraries should be pinned and reviewed before production use.

About

A responsible document revision tool for source-aware, personal, and transparent writing.

Topics

Resources

License

Stars

1 star

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors