Tavus offers white label conversational avatar capabilities to empower businesses to build their conversational avatars. 

Their avatars can do the following:

- Avatars can see passively and actively. They can recognize emotions and even view screenshare. 
- Avatars can adapt to how the end user talks. Tavus utilizes a mixture of semantics, conversational context and response time to determine when it's their turn to speak. 
- Avatars can demonstrate facial expressions and Tavus provides the sentiment analysis layer. 
- Avatars have latency as low as 600ms (response time)

Tavus also offers a white-label avatar creation API

- Tavus can finish cloning an avatar within 4 hours. Tavus just needs 2 minutes of training data
- Tavus moderates for you so you are compliant with AI ethics. Tavus will not clone anyone without their consent
- Tavus can clone AI characters as well

Tavus offers async video APIs

- Tavus offers lipsync, video script editing and text-to-video technologies that are very developer friendly. 
- Tavus' video script editing API and lipsync API are zeroshot
- 1 minute videos are generated within 3-5 minutes

Tavus offers the latest/greatest AI video models 

- Tavus' newest PHX3 model can do sentiment analysis and show facial expressions

Tavus is SOC2 compliant and HIPPA compliant.