Something we've been working on for a while, going out of stealth and open-sourcing it. With huge kudos to @RyanOnTheInside and the entire @DaydreamLiveAI gang.
There are many ways to describe this - you can even call it an AI instrument, but we'd love for you to go and check out the repo and play with it.
This was a lot of fun yesterday on @DaydreamLiveAI stream with @RyanOnTheInside playing with DEMON.
Check the recording below or just head to our YouTube to watch there https://t.co/4ccRuMzOpl
DEMON is AI music you can actually play like an instrument.
Yesterday we streamed a live walkthrough of DEMON with @RyanOnTheInside: our open-source real-time AI music engine for live remixing, source transformation, LoRAs, and MIDI-style control.
Full recording attached.
Details and the ultimate DEMON link collection in the thread. 🧵
There's now a paper on our DEMON engine on arXiv and a little video explainer.
What a week at @DaydreamLiveAI!
Go and play with it at https://t.co/Ed8dNqKRkn
To top off this week, there's now a paper about DEMON published on @arxiv.
A lot of great insights there - especially for researchers from the AI audio-gen domain.
Here's the video covering some basics, and the paper is in the tweet below.
Next week, on Tuesday (Jun 2), we are hosting our webinar with @RyanOnTheInside, author of the paper, so you can ask any questions.
Going from @suno → @DaydreamLiveAI to remix songs in real time is ridiculously fun.
The controllability and responsiveness are addictive. This very cool Suno track flips from lo-fi to metal in a heartbeat.
Ok, challenge accepted - made this on https://t.co/Ed8dNqKRkn for @emeisazam and @DaydreamLiveAI while waiting for GTA VI and November
Go give it a go, it's A LOT of fun!
🤠
Yesterday, we launched DEMON.
Today, we want to invite you to our DEMON live-streaming web event on June 2 at 9 AM PT / 12 PM ET / 6 PM CET.
@RyanOnTheInside will talk about DEMON, its internals, and use cases.
Join us!
@gootecks@RyanOnTheInside@DaydreamLiveAI It is a lot of fun, I love seeing folks paying with this, staring the repo - all that helps us a lot
https://t.co/Tw3IlCoBgA
Something we've been working on for a while, going out of stealth and open-sourcing it. With huge kudos to @RyanOnTheInside and the entire @DaydreamLiveAI gang.
There are many ways to describe this - you can even call it an AI instrument, but we'd love for you to go and check out the repo and play with it.
@IgorIlyinsky@RyanOnTheInside@DaydreamLiveAI You can play with it without MIDI controllers, too. MIDI is just making it easier to manipulate and turn those knobs.
You can try it out at https://t.co/eX0xKznCka
I don't know if I've missed your point entirely, though.
Introducing DEMON: a real-time, local-first diffusion engine for continuous music generation and control
Running at 25Hz with sub-100ms e2e latency on a single consumer GPU, it turns generative audio into an interactive streaming surface built to be played like an instrument.
I was animating some historical figures, but this can be way more useful for businesses across different verticals and use cases.
Here's the post covering build vs buy, sample project ideas/verticals where conversational, live AI avatars (running on your own stuff) deliver value: https://t.co/nmCnrVM7oy
Last week, we shipped the tutorial for a real-time conversational AI avatar in #Scope, running on a single GPU with a 24GB VRAM floor and LTX 2.3 model. You can see the demo of it running.
This week, we are taking it further and answering the question: When does running this locally beat a cloud API?
If you are making that decision, this is for you
🧵
Hey, it's time for a new demo, tutorial, and a new #Scope release.
The clip below is a real-time AI avatar built in Scope: you type a question, and it answers you on video.
The language model writing the reply, the @LTXStudio LTX 2.3 video model, and the audio all run together on a single GPU.
How we built it, plus a new Scope release, below. 🧵