\u26a0\ufe0f IMPORTANT DISCLAIMER: THIS SOFTWARE IS NOT A MEDICAL DEVICE. It is strictly an educational, technical, and research utility. It is NOT intended, designed, or cleared for medical diagnosis, clinical decision-making, patient care, or treatment. Always consult a qualified healthcare professional for medical advice.\n\nAI DICOM Reader is a specialized, privacy-first desktop utility designed for students, AI researchers, and tech enthusiasts. It bridges the gap between complex DICOM imaging formats and modern Artificial Intelligence Vision models.\n\nHave you ever wanted to test how well modern Large Language Models (LLMs) can interpret medical imaging data? AI DICOM Reader makes this possible by acting as a powerful local converter and viewer. It processes your DICOM files entirely offline, converts them into AI-friendly image formats, and allows you to securely prompt your favorite Vision AI models for educational insights.\n\nHow it works:\n1. Local Processing: Open DICOM files or folders directly on your PC. The app parses metadata, Hounsfield Units (HU), and pixel arrays completely offline.\n2. AI Integration: Connect your own API keys to query top-tier AI models, including OpenAI (GPT-4o), Anthropic (Claude), Google (Gemini), OpenRouter, or completely local, offline models via Ollama.\n3. Educational Summaries: The app compiles the AI's observations, your manual measurements, and DICOM metadata into structured, exportable educational reports.\n\nKey Capabilities:\n\u2022 Advanced Viewing: Apply Window/Level adjustments (Bone, Lung, Brain presets), pan, zoom, and utilize MPR (Multi-Planar Reconstruction) for 3D volume slicing.\n\u2022 Technical Measurement Tools: Draw regions of interest (ROI), measure distances, angles, and calculate basic pixel densities (HU/SI).\n\u2022 Privacy Controlled: No telemetry, no cloud syncing, and no hidden tracking. You control exactly what data is sent to your configured AI APIs. Local Ollama support ensures 100% offline AI analysis.\n\nNote: AI-generated observations are produced by third-party language models and may be factually incorrect, incomplete, or hallucinated. Use entirely at your own risk for learning and technical exploration.