Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Dialing in real time V2V in Sentinel. No Python or venvs, just an installer. Runs locally on any 16GB+ NVIDIA GPU. Sign up for early access:

Spencer Sterling

6,420 subscribers

58,438 views • 5 months ago •via X (Twitter)

Science & Technology

Anya Rossi• Live Now

Private livecam show

45 Comments

Sway Molina5 months ago

Cool! This reminds me of the stuff I experimented back with Runway Gen!

Artificially Inclined™5 months ago

Requested so gd hard I prob just broke you webhost

Adi Sideman5 months ago

Real-time V2V on a consumer GPU is the moment filmmaking stops being post-production and starts being live jazz. This is the instrument we've been waiting for 🎹

Sterling Crispin 🕊️5 months ago

Sick

MastaFrog5 months ago

Just an idiot and some c python.

Particular Uno5 months ago

How many System RAM needed? Looks amazing

Spencer Sterling5 months ago

16gb should be good! Everything is GPU accelerated, barely touches the CPU at all!

Particular Uno5 months ago

Great!

nin5 months ago

@Scobleizer Bout time!! 👊👊💪💪

Livepeer5 months ago

real-time is the only thing that matters for video tbh

B. Elliott Baker5 months ago

whoa!

PookieNumnums5 months ago

Spentinel

Infinite Vibes5 months ago

Exciting stuff man!

PXTCHXS5 months ago

Signed up! So cool!

Dan Moore5 months ago

@PurzBeats Gimmie

Mark Burke5 months ago

Great stuff...i'm new to this but find it fascinating. What are the prices please.

Abundance5 months ago

Nice

Bakka5 months ago

Is it goint to work on 3090?

Spencer Sterling5 months ago

It runs really well on a 3090!

Bakka5 months ago

Will try it Thanks

Truffle 🍄5 months ago

github pls

𝖇 𝖔 𝖔 𝖙 𝖊 𝖗5 months ago

@PurzBeats 👀

Syntax Diffusion5 months ago

🔥🔥

Slop-Swap5 months ago

Holy shit

Jakir Hossen5 months ago

this things is next level work

𝗡𝗢𝗥𝗔 💫🌸5 months ago

@neurocolor Nicely explain

Spencer Sterling5 months ago

@neurocolor This is a new real time graphics app i've been making that is able to achieve realtime warp-fusion and deforum style effects! It runs SDXL faster than any other implementation. It also has a modular vibe codable node system kinda like comfyui meets touchdesigner

SHOWLAB5 months ago

is it like StreamDiffusion?

Spencer Sterling5 months ago

Its similar, it doesn't use the multi-stream approach and instead focuses on single-step performance and real time graphics wizardry to tie it all together. Its fully implemented from scratch in C++ and runs about 40% faster than similar implementations

SHOWLAB5 months ago

thank you,really impressive,i saw Hugues Bruyère'spost here was tested with FLUX.2-Klein 4B, i did some testing with stable diffusion cpp but it can't support fp8 so fps not good as this

Spencer Sterling5 months ago

Thats dope! I did get klein running in real-time as well but never got it fully dialed in. I was able to quantize down to run with nvfp4 and get about 17fps, but the quality wasn't quite as good as I wanted since I am trying to optimize for 1 step generation

Professor Pixel5 months ago

@Scobleizer Holy crap that looks exciting

🪐 𝕄 𝕁 ✨5 months ago

lesssggooooo

nathaps5 months ago

@Scobleizer I’m on the list!

Johannes Stelzer5 months ago

@_pxlpshr sdxl turbo?

Spencer Sterling5 months ago

@_pxlpshr Yeah its SDXL Turbo, supports IP adapter, controlnet and v2v! Runs at up to 52fps on a 5090 at 896x512

Johannes Stelzer5 months ago

@_pxlpshr good old good old! I haven’t seen anything better for realtime stream -> video because its so so reactive…

Spencer Sterling5 months ago

@_pxlpshr Yeah, still the highest quality for the lowest latency as far as I know!

g0naji5 months ago

🔥

Max Fowler5 months ago

thias is so good ! , great work

Meta-Gnome5 months ago

Yes, So DOPE!

Doktor Funk5 months ago

This is dope

i wanna say thanks5 months ago

No need for Drugs

Zokkan5 months ago

only limitation the 16 Gigs Nvidia GPU xD

彡 𝔻█████𝕆 彡5 months ago

Amiga konnte das schon in den 90ern mit Picasso Karte.

Related Videos

Meet Liquid ShieldFlow. An on-device privacy layer powered by a device-native Liquid Foundation Model, Liquid ShieldFlow redacts sensitive data before it ever leaves your machine. No GPU needed and light on memory. It runs on almost any PC, locally, in real time. ShieldFlow was featured yesterday at Microsoft Build for Foundry Local. It also ran live on AMD laptops at 𝐂𝐎𝐌𝐏𝐔𝐓𝐄𝐗 𝐓𝐀𝐈𝐏𝐄𝐈. Request your early access here to ShieldFlow here:

Meet Liquid ShieldFlow. An on-device privacy layer powered by a device-native Liquid Foundation Model, Liquid ShieldFlow redacts sensitive data before it ever leaves your machine. No GPU needed and light on memory. It runs on almost any PC, locally, in real time. ShieldFlow was featured yesterday at Microsoft Build for Foundry Local. It also ran live on AMD laptops at 𝐂𝐎𝐌𝐏𝐔𝐓𝐄𝐗 𝐓𝐀𝐈𝐏𝐄𝐈. Request your early access here to ShieldFlow here:

Liquid AI

20,993 views • 3 months ago

Blender AI Motion Capture Plugin — Connects to a 1080P webcam and runs locally on your computer. Powered by a 1-billion-parameter model, it requires an 8GB VRAM GPU for real-time processing. Supports both real-time capture and video upload. The full-featured version currently supports NVIDIA CUDA only, and requires DX11 or above.

Blender AI Motion Capture Plugin — Connects to a 1080P webcam and runs locally on your computer. Powered by a 1-billion-parameter model, it requires an 8GB VRAM GPU for real-time processing. Supports both real-time capture and video upload. The full-featured version currently supports NVIDIA CUDA only, and requires DX11 or above.

CYANPUPPETS

58,611 views • 1 year ago

UE5.6 AI Real-Time Motion Generation Plugin — powered by a 1-billion-parameter motion model trained on NVIDIA CUDA. It runs locally in real time with 8GB of VRAM and converts webcam or video footage into real-time XYZ skeletal point data and rotation values.

UE5.6 AI Real-Time Motion Generation Plugin — powered by a 1-billion-parameter motion model trained on NVIDIA CUDA. It runs locally in real time with 8GB of VRAM and converts webcam or video footage into real-time XYZ skeletal point data and rotation values.

CYANPUPPETS

28,871 views • 8 months ago

UE5.6 AI Real-Time Motion Generation Plugin — powered by a 1-billion-parameter motion model trained on NVIDIA CUDA. It runs locally in real time with 8GB of VRAM and converts webcam or video footage into real-time XYZ skeletal point data and rotation values.

UE5.6 AI Real-Time Motion Generation Plugin — powered by a 1-billion-parameter motion model trained on NVIDIA CUDA. It runs locally in real time with 8GB of VRAM and converts webcam or video footage into real-time XYZ skeletal point data and rotation values.

CYANPUPPETS

15,968 views • 7 months ago

UE5.6 AI Real-Time Motion Generation Plugin — powered by a 1-billion-parameter motion model trained on NVIDIA CUDA. It runs locally in real time with 8GB of VRAM and converts webcam or video footage into real-time XYZ skeletal point data and rotation values.

UE5.6 AI Real-Time Motion Generation Plugin — powered by a 1-billion-parameter motion model trained on NVIDIA CUDA. It runs locally in real time with 8GB of VRAM and converts webcam or video footage into real-time XYZ skeletal point data and rotation values.

CYANPUPPETS

29,029 views • 8 months ago

Using glm 4.6 in Claude Code Locally on my PC for FREE. GPU: RTX 5080 16GB RAM: 64GB DDR5 5600MHz CPU: Intel Ultra 9 275HX 20 Cores

Using glm 4.6 in Claude Code Locally on my PC for FREE. GPU: RTX 5080 16GB RAM: 64GB DDR5 5600MHz CPU: Intel Ultra 9 275HX 20 Cores

Tushar Verma

81,072 views • 6 months ago

You don't need a GPU to run text-to-speech anymore. Kyutai released Pocket TTS, a 100 million parameter speech model that runs entirely on CPU. How it performs: > Nearly 6x faster than real time on a MacBook CPU > First audio chunk arrives in about 200 milliseconds > Runs on just 2 CPU cores > Supports voice cloning and 6 languages Setup takes one command: pip install pocket-tts Text-to-speech that used to need a GPU or a paid API now runs locally, on a laptop, in near real time.

You don't need a GPU to run text-to-speech anymore. Kyutai released Pocket TTS, a 100 million parameter speech model that runs entirely on CPU. How it performs: > Nearly 6x faster than real time on a MacBook CPU > First audio chunk arrives in about 200 milliseconds > Runs on just 2 CPU cores > Supports voice cloning and 6 languages Setup takes one command: pip install pocket-tts Text-to-speech that used to need a GPU or a paid API now runs locally, on a laptop, in near real time.

Simplifying AI

14,391 views • 1 month ago

someone open-sourced a tool that turns any phone video into a full 3D world in real time. you record a walk around your house, drop the mp4 in, and the entire scene comes back as a walkable 996k-point world at 60fps on your local GPU. no cloud. no subscription. runs 100% locally. 100% open source.

someone open-sourced a tool that turns any phone video into a full 3D world in real time. you record a walk around your house, drop the mp4 in, and the entire scene comes back as a walkable 996k-point world at 60fps on your local GPU. no cloud. no subscription. runs 100% locally. 100% open source.

Superman

819,309 views • 2 months ago

📣 Dev news incoming! Soon, you can experiment with an early preview of Android Studio on Project IDX running fully in your browser — no machine setup required. Sign up for early access program here →

📣 Dev news incoming! Soon, you can experiment with an early preview of Android Studio on Project IDX running fully in your browser — no machine setup required. Sign up for early access program here →

Android Studio

114,694 views • 2 years ago

"I don’t have a GPU" is no longer an excuse 🤯 You can now train LLMs directly in VS Code using a free Google Colab runtime. → Connect any fine-tuning notebook to Colab → Train locally or on a free cloud GPU → Works with Unsloth

"I don’t have a GPU" is no longer an excuse 🤯 You can now train LLMs directly in VS Code using a free Google Colab runtime. → Connect any fine-tuning notebook to Colab → Train locally or on a free cloud GPU → Works with Unsloth

Alvaro Cintas

86,321 views • 7 months ago

Our developer advocate Amos Gyamfi created a real-time AI voice agent that runs locally on your laptop. No GPU, no network calls. It uses Pocket TTS (100M params) for fast, natural speech paired with Stream Video. We're breaking it down ↓

Our developer advocate Amos Gyamfi created a real-time AI voice agent that runs locally on your laptop. No GPU, no network calls. It uses Pocket TTS (100M params) for fast, natural speech paired with Stream Video. We're breaking it down ↓

Stream

369,877 views • 7 months ago

Huge announcement TOMORROW. Join our mailing list for early bird access… Sign up now - link below! It’s MAKE OR BREAK time for Britain

Huge announcement TOMORROW. Join our mailing list for early bird access… Sign up now - link below! It’s MAKE OR BREAK time for Britain

Looking for Growth

37,740 views • 1 year ago

Blender AI Real-Time Motion Capture Plugin — connect a 1080P camera or upload videos. It runs locally with a 1-billion-parameter model and requires 8GB of VRAM for real-time processing. It supports both real-time capture and video uploads. The full-parameter version currently supports NVIDIA CUDA and requires DX11 or higher.

Blender AI Real-Time Motion Capture Plugin — connect a 1080P camera or upload videos. It runs locally with a 1-billion-parameter model and requires 8GB of VRAM for real-time processing. It supports both real-time capture and video uploads. The full-parameter version currently supports NVIDIA CUDA and requires DX11 or higher.

CYANPUPPETS

54,439 views • 6 months ago

Blender AI Real-Time Motion Capture Plugin — connect a 1080P camera or upload videos. It runs locally with a 1-billion-parameter model and requires 8GB of VRAM for real-time processing. It supports both real-time capture and video uploads. The full-parameter version currently supports NVIDIA CUDA and requires DX11 or higher.

Blender AI Real-Time Motion Capture Plugin — connect a 1080P camera or upload videos. It runs locally with a 1-billion-parameter model and requires 8GB of VRAM for real-time processing. It supports both real-time capture and video uploads. The full-parameter version currently supports NVIDIA CUDA and requires DX11 or higher.

CYANPUPPETS

26,492 views • 7 months ago

We just killed AI subscriptions One wallet. Every tool. No subscriptions. Just sign in. Early Access starts NOW

We just killed AI subscriptions One wallet. Every tool. No subscriptions. Just sign in. Early Access starts NOW

J

318,557 views • 1 year ago

RELEASE DATE REVEAL! Luminary's Early Access will launch July 30th A Solo or Co-op Openworld ARPG-Lite that focuses on straightforward intuitive gameplay. Let EVERYONE KNOW! (Links in Comments) Content Creator "Early Preview" Sign Up Week 1 Player's Name in Credits Sign Up

RELEASE DATE REVEAL! Luminary's Early Access will launch July 30th A Solo or Co-op Openworld ARPG-Lite that focuses on straightforward intuitive gameplay. Let EVERYONE KNOW! (Links in Comments) Content Creator "Early Preview" Sign Up Week 1 Player's Name in Credits Sign Up

Luminary | Refractive

182,279 views • 3 months ago