Conversational Replicas by Tavus - Build with real-time digital twins that speak, see & hear

With CVI, developers can now build live video conversations with digital twins – that actually feel real. It's the only end-to-end conversational pipeline with less than a second of latency. Check out a live demo on our website

Add a comment

Replies

Best
This has been the most most incredible product I've ever worked on!
it took the most incredible people to build it!!!
you’ve done some amazing work on the stock replicas and in discovering what makes a good one, as well as helping answer essentially research questions.
Thank you sooo much Brian!! It was truly the most exciting project I've worked on so far! Excited to continue learning and evolving alongside the Tavus crew <3
Wow! 🤯 Incredibly impressive, can't wait to see all the ways this will be used. Congrats on the launch!
Thank you Thijs- really appreciate the support 🙏
Thanks for check it out, and for the support!
Building this over the past few months has been an amazing experience - so proud of the whole team for the incredible effort to make this a reality! Can't wait to see what y'all do with this :)
great teams = great products.
great working with you directly on everything from latency to quality and everything in between. You really smashed everything you took on!
Researching and building models to make the conversational replica a reality has been an experience of a lifetime for me. To have the opportunity to work with some of the best technical minds made me pretty numb to the beauty of the product we have been building as we were interacting and building it brick by brick everyday. The magic moment of magnitude of the product we were building only struck me when I showed the conversational replica of myself (my digital twin) to my parents and my siblings, through a video call and asked them to interact with it. They were starstruck by the product we engineered, and I was too by their reactions.😅
Hah! I had the same experience- I was so deep in the weeds that I forgot how magical of an experience it was until I showed it to my mom the other day and her jaw dropped! 🤯
The new digital AI Era has started now.🤖🤖
Its only getting started!
This is so cool!! How do I get a few of these replicants on my team? Ngl the whole Tavus brand aesthetic is pretty sick too. Great work Tavus Team!!
Hey Carl!! We have a few different plans you can try out! If you need any further help - please let our team know!
So amped you were able to try it!!! And thanks for the compliment on the branding- has put it a lot of work there 🤩 Always excited to build together- please do reach out if we can be helpful 🚀
Appreciate the kind words.
Super exciting demo! I was initially skeptical but it's actually a nice experience to talk back and forth with an AI avatar who's very quick to respond. It wasn't pretending to be real - it's just a nicer interface than reading documentation or chatting, etc. Very curious to see where this goes!
Thanks Rajiv! Totally- we see this as a new interface to communicate with LLM and computers more naturally!
This is quite a cool demo of the digital twin tech. Quite impressive how realistic everything looks! (I think lips can improve - that's the only thing that gave away it's AI for me). Curious how do you ensure that users only create a digital twin of themselves? Props for thinking through all of the security issues that can arise! Congrats on the launch and team!
We do video consent verification for every Replica to ensure that Replicas can only be created by the person themselves.
- cool, that's what I thought.
The conversations looks pretty natural. Haven't seen any product in the market where I can talk to a replica like a zoom call. Being able to use just my voice to interact with the replica is a huge unlock (unlike other ones which only allow text input). The vision capabilities are a cherry on the top. Awesome product !!
The vision feature is my favorite- its so cool to incorporate into conversations. Amazing feedback, excited for you to keep using the product 🤩
This sounds really intriguing, ! The potential applications for the Conversational Video Interface seem endless. I'm particularly curious about the corporate training aspect; do you have any insights on how the interactive experience might enhance the learning outcomes compared to traditional methods? Also, with the example of expert cloning for customers like Delphi, how are you ensuring the authenticity and personality of the digital twin? I can see this tech being a game changer for industries like healthcare and education too. What do you think the biggest challenges will be in scaling this solution for broader use cases? Excited to see where this goes!
the number of use cases is amazing! For corporate training, we've seen these Conversational Replicas significantly increase engagement and retention at a F500 that runs a training program for new hires which was really cool For use cases like Delphi- there's some really cool optimizations that can be made with the LLM that drives the conversation! We're really excited to see how y'all continue to build with the product- please do share any feedback along the way 🦚
Measuring authenticity is definitely a work in progress and we don't have benchmarks as yet, but this is something that's on our radar. Right now it's a case of art and science, but our team is thinking deeply about it.