Phoenix-friendly wrapper for tesseract.js.

Run OCR in the browser — manga, document scanning, receipt parsing, anything with text in an image — without writing tesseract.js boilerplate. Pick a language, drop two components into your layout, and call getOcrWorker() from JS.

Quick start

# mix.exs
{:tesseract_js, "~> 0.1"}

# config/config.exs
config :tesseract_js, lang: "eng"

# root.html.heex
<TesseractJs.Component.preload />
<TesseractJs.Component.script />

// anywhere in your JS
const { getOcrWorker, recognize } = window.TesseractJs;
const { data } = await recognize(canvasOrImg);
console.log(data.text);

By default everything heavy loads from jsDelivr. To self-host, run:

mix tesseract_js.download eng jpn
# then in config.exs:
config :tesseract_js, source: :local

Modules

• TesseractJs.Models — language registry + URL builders.
• TesseractJs.Config — config defaults and runtime resolution.
• TesseractJs.Component — <.preload /> and <.script /> HEEx components.
• Mix.Tasks.TesseractJs.Download — mix tesseract_js.download task.