ጥበቡ is an Ethiopian tool build to recognize amharic text from images, OCRing them basically but built specifically for the amharic language.
# TIBEBU — ጥበቡ
> **Native Amharic OCR — Extract, read, and translate Ethiopic text from images.**
TIBEBU is a purpose-built optical character recognition tool for the Amharic language (አማርኛ). Powered by Google's Gemini 2.0 Flash vision model, it detects and extracts Amharic/Ethiopic text from images with high accuracy, and can optionally translate the result into English.
---
## Features
- **Drag & drop or click-to-upload** image input
- **Amharic text extraction** from photos, scans, screenshots, and documents
- **English translation** of extracted Amharic content
- **Clean, minimal dark-mode interface**
## Tech Stack
| Layer | Technology |
|-------|-----------|
| Frontend | React 19, TypeScript |
| Build | Vite 6 |
| AI Engine | Google Gemini 2.0 Flash (`@google/genai`) |
| Styling | Inline CSS (dark theme) |
## Getting Started
### Prerequisites
- Node.js 18+
- A Google AI Studio API key
### Setup
```bash
git clone
github.com
cd Tibebu-Amharic-OCR
cp .env.example .env
```
Add your Gemini API key to `.env`:
```env
GEMINI_API_KEY="your_key_here"
```
Install and run:
```bash
npm install
npm run dev
```
Open the address shown in your terminal (default `
localhost`).
### Production Build
```bash
npm run build
npm run preview
```
## Usage
1. Drop an image containing Amharic text onto the upload zone, or click to browse.
2. Click **Extract & Translate**.
3. View the extracted Amharic text alongside its English translation.
4. Use **Reset** to start over with a new image.
## Environment Variables
| Variable | Required | Description |
|----------|----------|-------------|
| `GEMINI_API_KEY` | Yes | Your Google Gemini API key |
## Why TIBEBU?
Most OCR tools treat Amharic as an afterthought. TIBEBU was built from the ground up for the Ge'ez script — designed by Ethiopians, for the Ethiopian language. The name *Tibebu* (ትበቡ) reflects excellence and distinction.