ocy 1.0.0

Lightweight OCR for developer screenshots — extract text from code images, terminals, and IDE windows

OCY - Python OCR SDK

Lightweight OCR for developer screenshots — extract text from code images, terminals, and IDE windows with one line of code.

Made by Samin Yeasar • github.com/solez-ai • x.com/Solez_None

Installation

pip install ocy

Quick Start

1. Set your API URL

First, set the API URL. You can use a self-hosted instance or a public one:

import ocy

# Set your API URL (required)
ocy.set_api_url("https://your-api.vercel.app")

2. Extract text from a URL

result = ocy.extract_text("https://example.com/screenshot.png")
print(result["text"])
print(f"Confidence: {result['confidence']:.1%}")

3. Extract text from a local file

result = ocy.extract_text("./myScreenshot.png")
print(result["text"])

4. Async usage

import asyncio
import ocy

async def main():
    result = await ocy.extract_text_async("https://example.com/code.png")
    print(result["text"])

asyncio.run(main())

5. Use an API key

import ocy

# Option 1: Per-request key
result = ocy.extract_text("https://example.com/screenshot.png", api_key="your-key")

# Option 2: Set global key
ocy.set_api_key("your-key")
result = ocy.extract_text("https://example.com/screenshot.png")

Response Format

{
    "text": "extracted text from image...",
    "confidence": 0.95,        # 0.0 to 1.0
    "latency_ms": 150,         # processing time in milliseconds
    "model": "ocy-v1-int8",    # model identifier
    "chars_detected": 42       # number of characters detected
}

API Reference

Function	Description
extract_text(img, api_key=None)	Extract text synchronously
extract_text_async(img, api_key=None)	Extract text asynchronously
set_api_url(url)	Set custom API endpoint
set_api_key(key)	Set global API key

License

MIT License - See LICENSE for details.