wtf is reactorx2fullworkflow?

This repository is not a piece of software you install and run. It is a worked example showing how to make a multi-character video using Reactor X2, a real-time video-to-video editing model built by a company called Reactor. Reactor X2 takes a live video source, such as a webcam feed, a video clip, or a still image, along with a written editing instruction and an optional reference image, and streams back a transformed version of that video as it plays. The repository bundles everything needed to follow along with one specific production: the reference character images used as identity input, the video clips X2 generated for each character, the background images used behind them, a file of reusable prompt recipes, and the final finished showcase video that resulted from putting it all together. To follow the workflow yourself, you open Reactor's own online playground, sign in, choose whether your source is a video or an image, upload one of the included reference images, copy the matching prompt from the prompts file, and start editing. A setting called keep frame backlog can be turned on for smoother continuity between frames, or off when faster response time matters more. The project also documents the exact step by step production sequence in a separate workflow file. For developers who want to build this into their own applications rather than using the web playground, Reactor also offers typed SDKs for JavaScript and Python, along with a command line tool to scaffold a new project pointed at the X2 model. The written material in this repository, such as the documentation files, is released under the MIT license. The character images and generated video clips themselves are excluded from that license, and a separate file explains the rules for using or sharing them.