Skip to content

Latest commit

 

History

History
30 lines (17 loc) · 935 Bytes

README.md

File metadata and controls

30 lines (17 loc) · 935 Bytes

WebSocket Transport Pipecat JavaScript Client SDK Example

Install and run locally

npm i
npm run dev

cp env.example .env
# now add API keys to .env

Open http://localhost:5173/

Demo code

This is bare-bones LLM voice chat app that can use either a Daily WebRTC connection or a WebSocket connection directly to the OpenAI Realtime API servers.

The code is intended as a proof-of-concept for implementing "transports" for the Pipecat RTVI (client-side) SDKs for Web, React, iOS, Android, and C++.

The application code is all in two files:

A very basic (not production-ready) transport implementation for the OpenAI Realtime API is in src/openai-websocket-transport.ts

The Pipecat RTVI (client-side) JavaScript SDK and the Daily WebRTC transport implementation are installed from npm.