RTVI Backend / Infrastructure Examples

This repository contains example applications for launching and deploying real-time bots that you can connect to with your RTVI-enabled clients. All of them will provide you with a base URL that you can pass to the Voice Client constructor, like so ...

import { VoiceClient } from "realtime-ai";

const voiceClient = new VoiceClient({
    baseUrl: "http://localhost:7860", // as an example
    enableMic: true,
    config: {...},
});

... and launch a bot that you can talk with.

Most of the examples use the Pipecat RTVI framework implementation for their bots. Where possible, we follow the design pattern established in the Pipecat documentation, which can be found here.

Available examples:

• local - A useful setup when building or testing clients in local development. This example spawns Pipecat bots that connect to your RTVI clients as local Python processes. It also serves as a good starting template for other deployment targets not included in this repository.

• Modal - Deploy bots to modal.com.

---

We would appreciate any pull requests for other cloud providers not included in this repository!