Tavus is the odd one out in this category: instead of generating a one-way scripted video, its core product is a live digital human that can see, hear, and respond in real time — built for developers embedding conversational agents into their own apps.
Tavus
by Tavus
Real-time, two-way conversational video agents — not a one-way scripted avatar tool.
- Named component models: Phoenix-4 (real-time facial rendering), Raven-1 (perception), Sparrow-1 (conversational turn-taking)
- Free Basic tier (25 min conversational video, 5 min generation, 25 stock replicas); Starter $59/mo, Growth $397/mo
- Digital 'replica' built from about two minutes of a person's own footage
- Supports 30+ languages; developer-focused via its Conversational Video Interface (CVI) API