AI model UI presentation

My own interface for AI models: a node graph where image, video, sound, 3D and language models work together, locally on my GPU or through online services, and where an AI "director" plans and builds the whole pipeline by itself. What the video shows: • The basics: nodes, wires, a model library that reads each model's own files (nothing guessed) • Images: local generation, consistent characters from reference pictures, live editing • Video: in-between frames from a first and last frame, frame-exact trimming • Audio: generated ambience with a live spectrum display • Long films: automatic shot detection, splitting into chunks, picking out one shot • Documents: a scanned 1959 book read page by page into clean text • 3D: animated scenes, cameras, keyframes and animation curves • Online services (Replicate, Google) next to local models, with the cost shown before anything runs and a budget per project • The Executor: describe what you want in plain language; the AI brain (Claude) proposes a plan, waits for approval, builds the project, checks every picture with a second model, and fixes what fails The example made inside it: Aesop's "The Tortoise and the Hare". 18 pictures, 9 animated shots, 3 minutes of narration, for about the cost of a lunch. Since then, the same approach has produced 64 illustrated folk tales from around the world, each narrated in 15 languages. Watch the tales: ‪@WorldTalesTold‬ Connect with me:   / aslan-zamaev-6a914145   Built by me, step by step, with the help of Claude Opus 5.5: Python, Qt, local and online AI models. Special thanks to Tamerlan Zamaev (cloud & software architecture) and Zulia Shavaeva (AI security) for their advice along the way.   / tamerlan-zamaev-b01a9a156     / zulia-shavaeva-72767113b