Catnip releases MaineCoon, a 22B streaming audio-visual model that generates real-time character video at 47.5 FPS

Original post

Robert Scoble@Scobleizer#447inTech

Other world models only let you play for a few minutes.

Someday soon I'll be able to send a script to one of these things and build a new kind of world/show/experience.

Is this it?

Love the name, signed up.

Catnip@catnips_ai

🥇MaineCoon: From Passive Video to Real-Time AI Presence

The first unlimited-duration interactive audio-visual model.

Most AI products today still feel like they live behind a screen.

You type. It answers. You speak. It replies.

The interaction is still mostly turn-based.

Mainecoon is built around a different idea: AI should not just respond to you. It should feel present with you.

🔗Learn more Website ↓ https://mainecoon.tech Blog ↓ https://mainecoon.tech/blogs

11:42 AM · Jun 16, 2026 · 4.4K Views

mainecoon.tech

MAINECOON.TECHVia

blogs

MAINECOON.TECH

VIEWS57KLIKES74RETWEETS15

Chubby♨️@kimmonismus

The interesting move with MaineCoon is that it doesn't stop perceiving once it starts responding.

Most conversational AI waits for input, answers, then goes quiet until you speak again; this one is built to keep reading expression, voice, and timing while it talks, in real time.

That shifts AI video from something you watch toward something that reacts to you while you're in it, and it's one of the clearer signals I've seen of where real-time, interactive AI is actually heading.

Catnip@catnips_ai

🥇MaineCoon: From Passive Video to Real-Time AI Presence

The first unlimited-duration interactive audio-visual model.

Most AI products today still feel like they live behind a screen.

You type. It answers. You speak. It replies.

The interaction is still mostly turn-based.

Mainecoon is built around a different idea: AI should not just respond to you. It should feel present with you.

🔗Learn more Website ↓ https://mainecoon.tech Blog ↓ https://mainecoon.tech/blogs

14h57K7427

BOOKMARKS29REPLIES25

Santiago@svpino

Look at the speed at which this model generates video!

Imagine a chat interface that uses real-time video streaming to talk to us instead of text.

We are getting close!

Catnip@catnips_ai

🥇MaineCoon: From Passive Video to Real-Time AI Presence

The first unlimited-duration interactive audio-visual model.

Most AI products today still feel like they live behind a screen.

You type. It answers. You speak. It replies.

The interaction is still mostly turn-based.

Mainecoon is built around a different idea: AI should not just respond to you. It should feel present with you.

🔗Learn more Website ↓ https://mainecoon.tech Blog ↓ https://mainecoon.tech/blogs

12h31.6K4329

Rohan Paul@rohanpaul_ai

Catnip just dropped MaineCoon, a 22B real-time audio-visual foundation model that turns text prompts into a live character stream with synced speech, motion, and expression.

The first streaming-native model of its kind.

sub-second first frame, 47.5FPS on one H100, 30FPS on one RTX Pro 6000, and about 7x faster throughput than comparable audio-visual systems in its internal tests.

The big deal is that a normal video generator can wait, revise, and render a finished clip, but a social interface has to move causally, remember its own imperfect past, and stay ahead of playback without breaking identity, voice, or rhythm.

Catnip@catnips_ai

🥇MaineCoon: From Passive Video to Real-Time AI Presence

The first unlimited-duration interactive audio-visual model.

Most AI products today still feel like they live behind a screen.

You type. It answers. You speak. It replies.

The interaction is still mostly turn-based.

Mainecoon is built around a different idea: AI should not just respond to you. It should feel present with you.

🔗Learn more Website ↓ https://mainecoon.tech Blog ↓ https://mainecoon.tech/blogs

1d3.2K136

Carlos E. Perez@IntuitMachine

Generation so fast it appears to be live:

15h5624

이상범@0thernes_ai

@Scobleizer https://0thernes.github.io/cosmogonic-quantum-mechalogodrom/ play with mine. (desktop monitor size is recommended)

1d211

Habibur Rahman@heyhabibur

@svpino This is closer to simulation than chat.

11h3421

Anna Stone@X5ZlJ

@svpino If AI feels continuously present, it changes emotional perception a lot.

11h321

Robert Scoble@Scobleizer

@0thernes_ai Cool, melted down my old iMac. Will get out my new MacBook to try again. But sure looked wild before everything stopped.

1d251

kent@kent350777

@svpino the speed demo always skips the interruption handling. real-time video chat means handling me cutting it off mid-sentence, which needs the stream to flush and re-sync state, not just generate fast.

11h211