Tavus

Overview

Tavus is a video generation and conversational video platform for building real-time AI replicas. Teams use its APIs and no-code tools to create digital humans that converse naturally, and to generate production-ready videos with AI presenters.

Key Features

• Custom replicas: Train a hyperrealistic AI replica of a person from about two minutes of video, voice model included.
• Conversational Video Interface: An end-to-end real-time pipeline with vision, turn-taking, and rendering models.
• Model stack: Phoenix-4.5 for rendering, Raven-1 for perception, Sparrow-2 for turn-taking, and Griffin in preview.
• Knowledge and memory: Ground conversations in uploaded documents or websites and carry context across sessions.
• Languages: Support for 42+ languages.

Use Cases

• AI sales development representatives and interviewers
• Healthcare and learning and development agents
• Personalized video messages at scale
• Real-time digital presenters inside meetings
• Interactive product demonstrations

Pricing

• Free: 20 minutes of conversational video and API access.
• Starter ($22/mo): 60 minutes, one custom face slot, and one concurrent session.
• Builder ($59/mo): 175 minutes, three custom faces, and pay-as-you-go overage.
• Growth ($397/mo): 1,300 minutes, seven custom faces, and up to ten concurrent sessions.
• Business ($975/mo): 4,000 minutes, 15 custom faces, and priority support.
• Enterprise: Custom pricing with white-labeling, compliance, and SLAs.