NPCs can talk back now!
Introducing NPCTalk: make the NPCs of any game talk in realtime. I made every NPC in GTA V answer to me in character, in their own cloned voice, in under 1s, and they act on what you say.
We packaged all the skills Claude learned into a single repo, built on open models: NVIDIA Nemotron ASR for transcription, Qwen 3.8 27B for realtime dialogue, and Fish Audio Drama 3 for the cloned voices, self-hosted on a serverless GPU.
https://t.co/fzj9Bwd2uh
Prompt to 3D character animation, now with high fidelity hand + finger tracking✌️
FLUX 3 Max video → WiLoR for hands + GVHMR for body tracking → any animated character!
open source: https://t.co/g9jauAsIcS
Introducing FLUX 3 Max, the latest post-trained video model from fal Research in partnership with @bfl_ai.
FLUX 3 Max is tied for #1 in image-to-video quality on independent third-party benchmarks, generating a 10-second, 720/768p video in ~12 seconds.
For the next three weeks, FLUX 3 Max is 75% off, making it the most affordable frontier video model at its quality level.
Prompt → animation at realtime speed 🕺
FLUX 3 Max generates the video, GVHMR pulls clean 3D motion from it on one GPU, and it's retargeted live onto any mixamo character. Fully automated flow!
Open source: https://t.co/g9jauAsIcS
We compared Haiku, Sonnet and Opus 5.5 models on a simple rigging & animation task! We tracked time taken and total token cost! We'll be releasing a more in depth benchmarking and evaluation for a range of headless blender tasks!
@gmartidesign none of them nailed it, they broke in opposite ways. haiku kept blender’s auto weights every run: smooth, but lift a leg and the chest moves 13cm. sonnet and opus caught the leaks and painted their own: zero bleed, but rubbery or pinched hips.
NPCs can talk back now!
Introducing NPCTalk: make the NPCs of any game talk in realtime. I made every NPC in GTA V answer to me in character, in their own cloned voice, in under 1s, and they act on what you say.
We packaged all the skills Claude learned into a single repo, built on open models: NVIDIA Nemotron ASR for transcription, Qwen 3.8 27B for realtime dialogue, and Fish Audio Drama 3 for the cloned voices, self-hosted on a serverless GPU.
https://t.co/fzj9Bwd2uh
NPCs can talk back now!
Introducing NPCTalk: make the NPCs of any game talk in realtime. I made every NPC in GTA V answer to me in character, in their own cloned voice, in under 1s, and they act on what you say.
We packaged all the skills Claude learned into a single repo, built on open models: NVIDIA Nemotron ASR for transcription, Qwen 3.8 27B for realtime dialogue, and Fish Audio Drama 3 for the cloned voices, self-hosted on a serverless GPU.
https://t.co/fzj9Bwd2uh
NPCs can talk back now!
Introducing NPCTalk: make the NPCs of any game talk in realtime. I made every NPC in GTA V answer to me in character, in their own cloned voice, in under 1s, and they act on what you say.
We packaged all the skills Claude learned into a single repo, built on open models: NVIDIA Nemotron ASR for transcription, Qwen 3.8 27B for realtime dialogue, and Fish Audio Drama 3 for the cloned voices, self-hosted on a serverless GPU.
https://t.co/fzj9Bwd2uh
coming soon to universal modder with multiplayer, for free! We'll also be open-sourcing the whole thing (just like any game we publish) so you can build anything you like on top of it
how our gopniks learned to squat:
1. ai video of a guy doing a slav squat
2. pull his skeleton out of the video (mediapipe, open-source)
3. Retarget it onto the slav warfare rigs in blender