Renoise Live
A near-real-time AI survival-show experiment where prompts steer characters and story beats while H3 Max Turbo generates the next clips.
The models and systems that generate the video stream. Browse the related AI Live Hub listings and check each project’s access requirements before opening it.
A focused doorway into AI Live Hub’s guide to channels, platforms, creator tools, and open-source broadcast projects.
A near-real-time AI survival-show experiment where prompts steer characters and story beats while H3 Max Turbo generates the next clips.
UNREEL
An open-source, self-hosted AI streaming service that writes scenes and renders MiniMax H3 Max Turbo video through fal as a channel plays.

An API for creating and manipulating live AI video streams, with documentation pointing to browser, TouchDesigner and OBS integrations.

A real-time video-editing API that transforms an incoming camera stream using text prompts and character reference images.
Daydream Scope
A workbench for running real-time generative video pipelines, with model selection, composable processing and WebRTC, NDI, Spout or Syphon output paths.
StreamDiffusionV2
An interactive video-diffusion system with released checkpoints, configurable inference and support for different GPU arrangements.
Live2Diff
A video-to-video research pipeline using one-way temporal attention to transform live input, with webcam and screen-video demos.
Self Forcing
Research code that trains video diffusion models on their own autoregressive rollouts and provides streaming-generation inference.
LongLive
A long-video research project whose 1.0 implementation supports sequential prompts and interactive generation, with a newer 2.0 infrastructure in the main repository.
Matrix-Game
A family of interactive world models, with the 3.0 release focused on streaming generation and longer-horizon memory.
SoulX-FlashHead
A streaming talking-head research system with released Lite and Pro model variants and a Gradio streaming mode.
SoulX-LiveAct
A human-animation research system for streaming interactions, with memory compression and controls for action or emotion.
Oasis 500M
The released smaller Oasis model and inference code generate gameplay frames conditioned on actions and an initial frame.

A public continuous AI-generated broadcast powered by FastH3, where scheduled segments set the scene and votes or prompts can influence what happens next.

A channel builder where an AI director writes each beat, a video model renders it with sound, and paid steering or tips can fund the next clips.

A public live gallery and creation flow that can start an AI-generated room from a prompt, image, or video, then preserve it as a replay or reopen it.

A multiplayer drawing-and-guessing game where H3 Max turns a player’s sketch into moving video while the room tries to identify the prompt.

A point-and-click adventure platform where a picture becomes a branching world: click an object, choose an outcome, and watch the transition play as film.

Odyssey’s multi-agent world-model preview lets multiple human or AI participants share and act inside the same generated simulation in real time.