# Real-time communication

StreamVC is a real-time low-latency voice conversion solution developed by Google. It is capable of matching the target voice's timbre while preserving the source voice content and prosody. This technology is particularly suitable for real-time communication scenarios such as phone and video conferences, and can also be used for applications such as voice anonymization. StreamVC achieves lightweight and high-quality voice synthesis through the architecture and training strategies of the SoundStream neural audio codec. It also demonstrates the effectiveness of learning the causality of soft speech units while providing whitening base frequency information to improve pitch stability without revealing the source voice timbre.

AQChatServer is a high-speed and convenient anonymous online real-time chat room powered by AI. Built on Netty and protobuf protocols for high performance, it is comparable to game backend development. The entire process does not require HTTP protocol, and supports the sending and receiving of text, images, files, audio, and video.

Muddy is a collaborative tool designed for teams. It simplifies workflows across multiple applications and documents using AI, allowing team members to collaborate more efficiently. Muddy can automatically organize and categorize tabs, supports unlimited undo functionality, enabling users to quickly switch between applications, files, and conversations. Additionally, it features universal commenting, allowing users to highlight, click, and send messages anywhere, similar to having Slack threads in every application and website. Muddy can also automatically read all tabs, learn from your conversations, and proactively ask follow-up questions when needed.

Rabbit is a research project aimed at developing a system that can understand and mimic human behavior in computer applications. This system, called the Large Action Model (LAM), uses neuro-symbolic programming technology, allowing for the direct simulation of various applications and users interacting with them. LAM rivals state-of-the-art methods in terms of accuracy, explainability, and speed. Its goal is to support the deployment of various AI assistants and operating systems, contributing to the shaping of the next generation of natural language-driven consumer experiences.