xirf

xirf / macula

Public

Lightweight OCR error detection and correction engine built in Zig

macula.andka.id ocr ocr-correction zig
21
3
100% credibility
Found Mar 02, 2026 at 21 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Zig
AI Summary

Macula is a lightweight browser tool that detects and corrects visual errors in OCR-scanned text, such as confusing similar-looking letters and numbers.

How It Works

1
👀 Discover Macula

You find a free online demo that fixes mistakes in scanned text from images.

2
🖥️ Open the Demo

The webpage loads quickly and the tool sets itself up without any waiting.

3
Add Your Text
📋
Paste Text

Copy messy scanned words from your OCR app and paste them in.

🖼️
Upload Image

Drop in a photo or screenshot and let it grab the text automatically.

4
🔍 Spot the Errors

Hit the button and watch it check every word for common mix-ups like 'rn' for 'm'.

5
📊 Review Fixes

A clear list shows good words, smart corrections, and any remaining issues.

✅ Perfect Text

You now have clean, readable text ready to copy and use anywhere.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 21 to 21 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is macula?

Macula spots and fixes common OCR mistakes—like "rn" for "m" or "0" for "O"—in text output from scanners or Tesseract. Built in Zig and compiled to WebAssembly, it runs entirely in the browser with a 800 KB payload that auto-loads on page open, no server needed. Paste OCR text into the live demo at macula.andka.id to see words flagged as valid, corrected, or errors, complete with confidence scores.

Why is it gaining traction?

This lightweight OCR library stands out for its tiny footprint and zero-dependency browser execution, beating bulkier Python models that need servers or heavy installs. Developers dig the instant demo, configurable threshold via JS, and seamless integration as a Zig lib or WASM module—perfect for real-time correction without latency. It's the best lightweight OCR model for client-side apps where every KB counts.

Who should use this?

Frontend devs embedding OCR in web apps like document scanners or photo-to-text tools. Indie hackers building offline mobile hybrids with Tesseract.js. Anyone processing scanned PDFs or screenshots needing quick error cleanup without cloud APIs.

Verdict

Grab it if you need a lightweight OCR GitHub gem for browser use—the demo and Zig API make prototyping fast, with solid tests covering the pipeline. At 12 stars and 1.0% credibility, it's early but mature enough for experiments; rebuild the model artifact yourself for custom dicts.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.