🎬 Video Analysis
💡Tips on Using Analysis Modes
⚡
Fast Analysis — Two-stage screening (structured video description via Gemini Vision, then clinical interpretation via Gemini Flash). Provides a concise conclusion and risk signal.
🧠
AI Expert Validated — Most accurate expert analysis (Gemini JSON + Opus 4.6) — recommended for detailed clinical review of video materials; most resource-intensive mode.
•🛡️ Video is automatically processed: the system extracts 5-12 key frames (adaptive to video length).
•🔒 Each frame is anonymized: black bars on edges (10% top, 8% bottom, 12% sides).
•👁️ Preview before sending: you can see all frames and manually edit any of them.
•🎯 ~92-95% accuracy with 7 frames — optimal quality and safety balance.
•💰 5-6x cost savings compared to sending the full video.
•⏱️ Processing takes 4-8 seconds depending on video length.
⚠️ Service availability notice
Some AI providers may be temporarily unavailable due to regional restrictions and provider routing. Users outside Russia may still use Claude, but availability is not guaranteed in every region. If Claude is unavailable, switch to GPT-5.2 in Optimized mode.
Upload Video for Analysis
📁
or
Supported: Video (MP4, MOV, AVI) or DICOM series (folder)
Patient database is empty. Add patients in the «Patient Database» section.