LLMOCR

LLMOCR uses a local LLM to read text from images.

You can also change the instruction to have the LLM use the image in the way that you prompt.

Features

• Local Processing: All processing is done locally on your machine.
• User-Friendly GUI: Includes a GUI. Relies on Koboldcpp, a single executable, for all AI functionality.
• GPU Acceleration: Will use Apple Metal, Nvidia CUDA, or AMD (Vulkan) hardware if available to greatly speed inference.
• Cross-Platform: Supports Windows, macOS ARM, and Linux.

Installation

Prerequisites

• Python 3.8 or higher
• KoboldCPP

Windows Installation

1. Clone the repository or download the ZIP file and extract it.

2. Install Python for Windows.

3. Download KoboldCPP.exe and place it in the LLMOCR folder. If it is not named KoboldCPP.exe, rename it to KoboldCPP.exe

4. If you want the script to download a model for you and have KoboldCpp run it for you, open llm_ocr.bat

5. If you want to load your own model using KoboldCpp, open llm_ocr_no_kobold.bat

Mac and Linux Installation

1. Clone the repository or download and extract the ZIP file.

2. Install Python 3.8 or higher if not already installed.

3. Create a new python env and install the requirements.txt.

4. Run kobold with flag --config llm-ocr.kcppt

5. Wait until the model weights finish downloading and the terminal window says Please connect to custom endpoint at http://localhost:5001

6. Run llm-ocr-gui.py using Python.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgements

• KoboldCPP for local AI processing
• PyQt6 for the GUI framework