The SDK exposes APIs and React components to control:

- Audio connection state - Real-time transcription - Conversation context and memory - Agent lifecycle and user messaging

This gives frontend teams full flexibility to implement multimodal AI workflows with low-latency audio and reliable WebRTC transport.

Separately, Telnyx has added image upload support to its default widget, which works with vision-capable LLMs like GPT-4o and LLaMA 4 Maverick (via Groq).

Learn more at https://telnyx.com/release-notes/widget-library or reach out to [email protected] with any questions.