AnyTalker Demo

Experience AnyTalker's multi-person talking video generation capabilities. This demo showcases how the framework creates natural conversations between multiple speakers, all driven by audio input.

How AnyTalker Works

AnyTalker generates talking videos featuring multiple people simultaneously. The framework uses an extensible multi-stream processing architecture with identity-aware attention to coordinate the movements and expressions of all speakers, creating believable multi-person scenes.

Key Features Demonstrated

  • Multi-Person Generation: Animate multiple speakers in the same scene
  • Audio-Driven Animation: Drive each person with their own audio track
  • Lip Synchronization: Accurate mouth movements matching the audio
  • Natural Interactions: Appropriate social behaviors between speakers
  • Identity Scalability: Support for arbitrary number of speakers
  • High Visual Quality: Realistic facial movements and expressions

Demo Capabilities

Two-Person Conversations

Watch as AnyTalker generates natural conversations between two people. Each speaker has their own audio input, and the system coordinates their facial animations to create believable interactions. Notice how the speakers display appropriate listening behaviors when the other person is talking.

Group Discussions

Experience multi-person group discussions with three or more speakers. AnyTalker handles multiple simultaneous speakers, maintaining lip synchronization for all participants while displaying natural social dynamics. The identity-aware attention mechanism ensures each person's animation is coordinated with others in the scene.

Interactive Scenarios

See how AnyTalker generates videos with natural interactivity between speakers. The system refines social dynamics through its training process, enabling speakers to make eye contact, display appropriate reactions, and exhibit natural conversational behaviors.

Video Demonstrations

Watch video demonstrations of AnyTalker in action. These examples showcase the framework's capabilities in generating multi-person talking videos with natural interactions and high visual quality.

Example 1: Two-Person Conversation

This demo shows two people engaged in a natural conversation. Notice the lip synchronization for both speakers and how they display appropriate listening behaviors.

Video demonstration placeholder

Example 2: Multi-Person Discussion

Watch a group discussion with three speakers. The framework handles multiple simultaneous speakers while maintaining natural social dynamics.

Video demonstration placeholder

Example 3: Interactive Scenario

See how speakers display natural interactivity, including eye contact and reactions to each other's speech.

Video demonstration placeholder

Technical Details

AnyTalker uses several key technologies to achieve natural multi-person video generation:

  • Identity-Aware Attention: Distinguishes between different people in the scene
  • Multi-Stream Processing: Handles multiple identity-audio pairs through parallel streams
  • Diffusion Transformer: Provides the foundation for high-quality video generation
  • Interactivity Refinement: Ensures natural social dynamics between speakers
  • Efficient Training: Learns from single-person videos and refines with few multi-person clips

Use Cases

AnyTalker's multi-person video generation capabilities enable various applications:

Content Creation

Generate engaging multi-person video content for social media, educational materials, or entertainment without filming.

Virtual Meetings

Create synthetic participants for demonstrations or prototypes of virtual meeting technologies.

Education and Training

Develop interactive educational content featuring multiple instructors or characters with realistic conversations.

Gaming and Animation

Animate multiple characters in games or animated content using voice acting as input.

Learn More

To learn more about AnyTalker's technology and capabilities, explore the following resources:

  • Read the research paper on arXiv (arXiv:2511.23475)
  • Visit the official project homepage at HKUST-C4G
  • Watch detailed video demonstrations on YouTube
  • Explore the technical documentation on the project website

Note: This demo page showcases AnyTalker's capabilities. For access to the actual demo system or research code, please refer to the official project repository and homepage.