Hi OpenAI team,
We’re building a ChatGPT app with an interactive widget, and we have a question about voice interaction support.
In the ChatGPT app, there is a voice conversation button next to the message composer. However, when users start a voice conversation from there, it does not appear to trigger our app tools. My understanding is that Voice Mode does not currently support ChatGPT apps/tools yet.
A few questions:
- Is there any general roadmap or expected direction for when Voice Mode may support ChatGPT apps and tool calls?
- Many of our customers strongly prefer voice-based conversations, especially for high-touch workflows where speaking feels more natural than typing.
- Would it be aligned with the ChatGPT Apps / Apps SDK development model if we add our own voice conversation button inside our widget, then use that to let users speak directly with our experience?
- If that approach is acceptable, are there any UX, safety, permission, or review guidelines we should follow to make sure the experience feels native to ChatGPT and does not conflict with the platform’s voice interaction model?
Our goal is not to work around the ChatGPT native voice experience, but to provide a voice-first interaction path for users until native Voice Mode can support apps/tools directly.
Any guidance on the recommended implementation pattern would be greatly appreciated.
Thanks!