Integration

OpenAI Realtime embed

Basic coffee-barista embed with OpenAI Realtime as speech-to-speech — pipe PCM + transcript into createUtterance.

Experience
OpenAI Realtime
createUtterance
BYO voice
transcript

Live demo Run locally GitHub source Experience on Meet spec.md

What it demonstrates

Hello-world embed — one experience id and one Experience / web component mount.
Public session mint with conversation-mode player defaults.
No host-side speak, listen, or lesson chrome.

Implementations

SvelteKit

`<Experience />` from `@liforma/client/svelte` — full lesson UI with TypeScript and Svelte 5.

View implementation →Vanilla HTML

Vanilla HTML

CDN script + `<liforma-experience>` — copy-paste friendly for any site.

View implementation →Next.js

Next.js

`@liforma/client/react` plus optional App Router session route helper.

View implementation →React (Vite)

React (Vite)

Client-side React with `@liforma/client/react` — no Next.js server.

View implementation →

Related

AI agent prompt

Use the Liforma OpenAI Realtime embed example to build an app in your chosen framework.

Read:
- https://github.com/LiformaLtd/examples.liforma.ai/blob/main/examples/openai-realtime-embed/spec.md
- https://github.com/LiformaLtd/examples.liforma.ai/tree/main/examples/openai-realtime-embed

Preserve the hello-world Experience embed (one experience id, no lesson catalogue).