tavus-video-gen

Quick start guide for the Tavus Conversational Video Interface (CVI) API. Covers the minimal setup to launch a real-time AI video conversation using stock resources or custom personas.

Controls Tavus CVI (Conversational Video Interface) conversations in real-time through the Interactions Protocol, enabling text echo, interrupts, context injection, tool call handling, and event listening via WebRTC and Daily.js.

An AI-native security testing platform built in Go that integrates 100+ security tools, an intelligent orchestration engine, role-based testing, a skills system, and a built-in C2 framework, all accessible through native MCP protocol and AI agents for end-to-end vulnerability discovery.

Provides pre-built React components and hooks for embedding Tavus CVI (Conversational Video Interface) into web applications. Includes a CVIProvider wrapper, Conversation component, and support for Vite and Next.js setups.

A skill for creating and managing Tavus replicas (AI digital twins) for video generation and real-time conversations, covering training video requirements, consent statements, and the Phoenix-3 model.

An overview skill for Tavus, an AI research lab building the Conversational Video Interface (CVI) for real-time, face-to-face AI conversations. Covers the Phoenix/Raven/Sparrow model stack, persona configuration, and WebRTC-based sessions with ~600ms latency.

A skill for configuring Tavus CVI (Conversational Video Interface) personas with custom LLMs, TTS engines, speech-to-text, perception via Raven, and turn-taking settings via Sparrow. Supports built-in models and bring-your-own-LLM through any OpenAI-compatible API.