Language, images and voice.
Generate copy, images and speech through the platform's audited OpenAI sidecar — the app never holds the key.