Screen-capture OCR tool with AI-powered text analysis, built in Java.

Neolithic is a desktop utility that lets you capture any region of your screen, extract text from it using Tesseract OCR, and then optionally send that text to GPT-4 for analysis, summarization, or Q&A — all without leaving your workflow.

• Global hotkey capture — Press Ctrl + Shift from any application to trigger the capture overlay
• Freehand area selection — Click and drag to select exactly the region you want
• Live preview — See the selected area before confirming
• Tesseract OCR — Accurate text extraction from screenshots using the bundled eng.traineddata model
• GPT-4 integration — Send extracted text to OpenAI for analysis with a custom prompt
• Interactive chat — Continue the conversation with the AI in a built-in chat window
• Image preprocessing — Converts and processes captured images for better OCR accuracy

• Java 17
• Swing (desktop UI)
• Tesseract / Tess4J (OCR)
• OpenAI Java SDK (GPT-4 API)
• JNativeHook (global keyboard shortcuts)
• Maven

• Java 17+
• Maven 3.8+
• An OpenAI API key

Option 1: Environment variable (recommended)

Set OPENAI_API_KEY in your environment:

# Linux/macOS
export OPENAI_API_KEY=sk-your-key-here

# Windows
set OPENAI_API_KEY=sk-your-key-here

Option 2: Config file

cp src/main/resources/config.example.json src/main/resources/config.json
# Edit config.json and add your API key

⚠️ Never commit config.json — it's excluded by .gitignore.

mvn clean install
mvn exec:java -Dexec.mainClass="org.codered.neolithic.Neolithic"

1. Launch the application
2. Press Ctrl + Shift anywhere on your screen
3. Click and drag to select the area you want to capture
4. Review the preview — click Accept to proceed
5. Neolithic runs OCR on the selection and opens a chat window
6. The extracted text is automatically sent to GPT-4 with a default prompt
7. Continue the conversation to ask follow-up questions about the captured content

src/main/java/org/codered/neolithic/
├── customize/          # Configuration management
├── images/             # OCR conversion, image preprocessing, UI
│   ├── conversion/     # Image-to-text pipeline
│   ├── processing/     # Image manipulation utilities
│   └── ui/             # Conversion result UI
├── openai/             # OpenAI API handler and request models
├── screenshot/         # Screen capture tool and hotkey listener
├── utils/              # Config reader (env + JSON)
└── Neolithic.java      # Application entry point

Open-source under the MIT license.