MemoryVerse AI '26

Setup 1. Install Ollama (the local LLM runtime) Windows / Mac: Download and run the installer from https://ollama.com/download Linux: curl -fsSL https://ollama.com/install.sh | sh Verify it's installed: ollama --version 2. Pull the model (one-time, ~1GB download) ollama pull qwen2.5:1.5b This model was picked because it's small enough to run comfortably on 8GB RAM machines while still being solid at structured JSON extraction. If your machine has more headroom later, llama3.2:3b gives noticeably better extraction quality — just ollama pull llama3.2:3b and change MODEL_NAME in extraction.py. 3. Start Ollama (it usually auto-starts as a background service; if not:) ollama serve Leave this running in a terminal. It serves the local API at http://localhost:11434. 4. Install Python dependencies cd memoryverse python -m venv venv source venv/bin/activate # Windows: venv\Scripts\activate pip install -r requirements.txt OCR note: for scanned certificate images, pytesseract needs the Tesseract binary installed separately: Mac: brew install tesseract Linux: sudo apt-get install tesseract-ocr Windows: installer at https://github.com/UB-Mannheim/tesseract/wiki If you skip this, everything still works for PDFs/docx — just scanned image certs won't OCR. 5. Run the app streamlit run app.py Opens at http://localhost:8501.