an american accent suits your vibe, she makes a great rep for you Paul Scanlon!

We're excited about disentanglement! 🧶➡️

...huh?

Here's commentary from our senior AI researcher Susho, PhD on last week's Microsoft VASA-1 paper.

He thinks one of the key breakthroughs is being able to cleanly disentangle the image into four independent buckets👇

1. 🤡 3D

Shout out to CapCut which we used with our AI stock replicas to create this quick video.

🎞️Generated videos from text using our APIs
✂️Then, we put the video together in Cap Cut

What other video editing tools should we try?

Our customers are building some cool apps and features using our Replica APIs, here's a little taste.

AI Replicas =

👩‍🎨 Cost effective AI influencers for marketing
🧑‍💼 Digital twins to make sales folks more scalable
🏢 Automated relationship management for realtors
💻 Marketing

Our eng team has been working around the clock to enhance our replica and text-to-video API performance.

Some improvements since our launch last month:

⌛️Reduced replica training time by 35%
📻Improved naturalness of audio
🪨Significant upgrades in model stability & robustness

We're excited to welcome Sushobhan Ghosh to the team!

Susho, PhD joins us as a Senior AI Researcher in San Francisco. He'll be focused on building robust and seamless custom replicas for our customers and developers.

Susho joins us from Descript where he worked on transcript

Part of making a digital replica involves reconstructing 2D footage into a 3D model.

Here is an example the 3D output we see under the hood 🔎

Our model took our Eng Manager's 2D training footage, and reconstructed his face with visualized mesh.

Neural radiance fields take

OH on LinkedIn on our Replica:

'Interesting, a lot of AI clones are out there. This one seems to be the most convincing.'

'Honestly this one has blown my mind! I do quite a few videos and this one helps a lot.'

Good luck to all the Y Combinator companies on demo day!

...video generated by Tavus featuring Hassaan Raza's replica 😎

25+ AI Companies from Y Combinator that have Trained their Own AI Models Instead of Using Someone Else’s Closed Model Through an API like a Black Box

marktechpost.com/2024/04/02/25-…

Atmo Deepgram Diffuse Bio Draftaid Navier AI (W24) metalware Linum Lina Colucci, PhD

“We see a future where everyone wants to have a digital replica of themselves; they control that and they have full authority over that,” Raza said.

techcrunch.com/2024/03/12/gen…

#avatar #replica #texttovideo #APIs

You can now build with our APIs too 🤓

Do your users make talking-head videos? Would they prefer to generate video rather than record themselves?

HubSpot is taking personalization to the next level with Tavus!

Instead of slapping [First Name] into emails, their team will be sending tailored genAI video messages.

You can also elevate your customer connections.
Discover more news and tips:
🔗crm.org/crm-newsletter

Tavus x Times Square

Our NYC team got the selfie 😎🗽

Build talking-head AI videos: tavus.io/developer

Happy to be up there with our friends Brex!

Mom, look! We're on a billboard in Times Square.

Get our replica & text-to-video API at tavus.io/developer

And make human-like videos without recording.

#GrowWithBrex