What is Riverst?

Riverst is an open-source research platform that makes it easy to create, deploy, and study interactive user–avatar conversations powered by AI and real-time voice.

Build

Configure multi-modal AI pipelines with speech-to-text, language models, and text-to-speech. Define custom system prompts, activity flows, and avatar behaviors — no ML expertise required.

Run

Deploy avatar interactions over WebRTC with real-time lip-sync, body animations, and low-latency audio. Works on CPU or GPU, in the cloud or on-premise.

Analyze

Every session is automatically logged, transcribed, and summarized. Review conversation analytics, download session data, and iterate on your activity design.

Research Background

Activities

Riverst ships with ready-to-use activity templates. Each activity configures the avatar's personality, goals, and conversation flow.

See Riverst in Action

Watch Riverst in action, then try it yourself with a live avatar session.

Watch the Demo

Live Demo

An interactive demo is coming soon. Watch the videos above to see Riverst in action.

Live Demo Coming Soon

The interactive demo is currently in progress. Watch our videos below to see Riverst in action.

Watch Videos

The Team

Deploy Your Own Instance

Riverst is fully open source. Run it on your own infrastructure in minutes.

  1. Clone the repository

    git clone https://github.com/sensein/riverst.git
cd riverst

  2. Configure environment variables

    Copy the example env files and fill in your API keys:

    cp src/server/.env.example src/server/.env
cp src/client/react/.env.example src/client/react/.env

    Required variables: OPENAI_API_KEY, GOOGLE_CLIENT_ID, GOOGLE_CLIENT_SECRET

  3. Launch with Docker Compose

    Then open http://localhost:3000 in your browser.

View on GitHub Report a Bug or Request a Feature Contributing Guide