Anva is the first CPU based real-time live avatar API. Reducing cost up to 10x compared to competitors such as Anam.ai and Tavus. We support BYO LLM / voice -> Where integrating Anva does not disturb your original intelligence logic. (We handle expressions / emotions automatically from context). Applications include: Language learning tutor, AI interview, Landing page inbound / sales, Coach / tutor, Virtual meetings, museums + education. And anything you can think of!
Hey guys,
I've been building avatar systems since 2019, back then with GPT-2 and so on. I've been iterating on Anva as a hobby and had at least 4 generations built, each improving on the other, over the years.
I'm finally happy to release the first public version of Anva (V4.9) and excited to work with anyone who might find integrating live avatars useful in their product, or create entirely new products out of it :)
More updates coming soon... including enhanced realism and full body / gestures!