LLM OCR Studio

This is a modern web interface for OCR using Vision LLMs.

🚀 Quick Start

1. Install dependencies:

npm install
npm run dev

Open your browser at http://localhost:3000.

2. Configure LLM

The application runs entirely in the browser but needs an OpenAI-compatible API to process images. The interface will warn you if the configuration is missing.

Click the Settings icon in the top right to configure:

• LM Studio (Local):
  • Base URL: http://localhost:1234/v1
  • Model: Enter the ID of the model loaded in LM Studio (e.g., llama-3.2-vision).
  • API Key: Leave empty.
• OpenRouter: https://openrouter.ai/api/v1 + API Key + Model Name.
• Any OpenAI Compatible Endpoint.

3. Usage

1. Trascina un PDF nell'area di upload.
2. Puoi trascinare più file o aggiungerne altri nella schermata di revisione (Batch processing).
3. The app converts files to images (Client-side).
4. Sends images to the configured Vision LLM.
5. Returns formatted Markdown.

🛠 Requisiti di Sistema

• Node.js: v18+
• API Access: Access to a Vision model (e.g., GPT-4o, Gemini 1.5, Qwen-VL via LMStudio).