what is comfyui_bernini_director fr?

This project is a custom node for ComfyUI, a visual workflow tool for AI image and video generation, built on top of the official Bernini video model. It packages timeline splitting, prompt encoding, two stage sampling, and export into a single director style node aimed at generating longer videos made of multiple segments rather than one short clip. Inside the node you upload a source video and can split it into segments manually, evenly, or automatically using scene detection, with a visual timeline showing each segment's range. The node supports many task modes, covering plain text to image, text to video, image to image, image to video, video to video, and more specialized modes such as inserting an advertisement into a video or making small positional tweaks to a subject already in a video. It can take up to five reference images that prompts can point to directly, and it runs two separate sampling passes, called HIGH and LOW, each with its own settings for guidance strength, seed, and step count. An optional cross segment guidance mode can be turned on to keep multiple segments visually consistent, and a built in connection to large language models can expand your prompts using Bernini's own prompt template. The node outputs the generated video frames, an audio track when the source video had one, frame rate and frame count information, and a text report summarizing the segment plan. To use it, you need a version of ComfyUI that supports the official Bernini model, then install this node by cloning it into ComfyUI's custom_nodes folder or through the ComfyUI Manager. The corresponding Bernini model weights and example workflows are distributed separately through the author's own site. The project is released under the Apache-2.0 license.