Category · live channels
AI live streaming — the audience-directed wave
AI live streaming — the realtime, audience-directed kind — arrived as a working format in late August 2026. Instead of rendering a video and uploading it, these channels generate each moment live, and the viewer (or the chat) effectively becomes the director: you prompt or vote what happens next, and the winning choice renders in seconds.
收录 · 6
由聊天区决定“接下来播什么”的实时生成视频直播。
fal.live
“人人皆可导演的 AI 电视”——观看生成式直播,并为“接下来发生什么”投票。基于 fal 运行。fal.live
InfiniteSlop
24/7 由聊天驱动的无限 AI 视频——你发提示词,几秒内便生成并持续播放。作者 Pieter Levels,运行于 fal。infiniteslop.ai
Renoise Live
永不停机的生存直播——四个岛民的命运,由实时聊天决定。renoise.live
Infinite TV
Open-source, Twitch-chat-driven AI TV — messages become prompts, LTX Video renders clips live, frames stream back over RTMP.github.com
SlopTV
Open-source “infinite AI slop generator from YouTube comments” — your live chat becomes clips, streamed back into the same broadcast.github.com
PixVerse R1
A realtime world model with shared live worlds — a room of people type prompts and the world renders instantly for everyone.realtime.pixverse.aiTwo formats, one substrate
fal.live runs curated channels ("AI television by everyone") where viewers pick from live candidate next-steps. InfiniteSlop is the wilder sibling — an open prompt-fed queue where anyone steers and content is uncurated by design. Both sit on fal’s realtime H3 Max build of MiniMax H3, and both are early: coverage still calls them showcases rather than finished products.
常见问题
What is AI live streaming?
A livestream where AI generates the video in real time, fast enough that the audience can steer what happens next — unlike a rendered-then-uploaded AI video.
What channels can I watch?
fal.live (curated AI TV channels) and InfiniteSlop (an open chat-directed stream) are the two live reference points as of Sep 2026.
Is AI live streaming free?
Viewing appears free on both; fal.live had published no pricing as of Sep 2026, while InfiniteSlop’s inference cost is sponsored by fal (~$4k/day, reported).
What makes realtime AI streaming possible?
Faster-than-playback generation — a model that produces roughly 5s of video in under 3s — which is what fal’s H3 Max build of MiniMax H3 provides.