OCR Studio

A beautiful, Cursor-inspired OCR web application powered by Z.AI's GLM-OCR.

Features

• Upload images (PNG, JPG) or PDFs
• Extract text using state-of-the-art OCR
• Edit extracted text in a code editor-style interface
• Download results as Markdown
• Copy to clipboard

File Limits and Large File Handling

GLM-OCR accepts single image files up to 10MB and PDF files up to 50MB.

This app adds client-side preprocessing so larger uploads can still be processed:

• Oversized images are automatically compressed before OCR.
• Oversized or long PDFs are split into OCR-safe chunks and processed sequentially.
• Chunk results are merged back into a single Markdown output in order.

Note: if an individual PDF page is too large to fit under API limits even by itself, the request will still fail and the file must be reduced manually.

Setup

1. Clone the repository
2. Install dependencies:

   npm install

3. Create a .env.local file with your Z.AI API key:

   ZAI_API_KEY=your-api-key-here
   

4. Run the development server:

   npm run dev

Deploy to Vercel

1. Push your code to GitHub
2. Import the project to Vercel
3. Add the ZAI_API_KEY environment variable in Vercel settings
4. Deploy

Tech Stack

• Next.js 14
• TypeScript
• Tailwind CSS
• Z.AI GLM-OCR API