{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreig44vg44azx7nju2agjayhmvfq7djkfhosbiw2j4ospjyo4jklrx4",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mpqshdl7xme2"
  },
  "path": "/t/microphone-input-and-audio-media-stream-output-support/1385656#post_1",
  "publishedAt": "2026-07-03T14:42:01.000Z",
  "site": "https://community.openai.com",
  "textContent": "Hi everyone,\n\nI’m building a ChatGPT app with an Apps SDK widget, and I want to confirm the expected support for media I/O inside the widget.\n\nThis is not about using ChatGPT’s native Voice Mode or triggering MCP tools from Voice Mode. I’m considering implementing voice directly inside the widget:\n\n  * capture microphone input with getUserMedia()\n  * connect from the widget to my backend using WebRTC or WebSocket\n  * receive and play back audio/media streams inside the widget\n\n\n\nQuestions:\n\n  1. Is microphone/media capture inside an Apps SDK widget officially supported or allowed if the user grants permission?\n  2. Is audio playback or WebRTC media output expected to work inside the widget iframe?\n  3. Are there known limitations or bugs on ChatGPT native mobile apps, especially iOS/Android, compared with ChatGPT web/desktop?\n  4. If this is allowed, are there recommended CSP, permission, or implementation patterns for this use case?\n\n\n\nThanks!",
  "title": "Microphone input and audio/media stream output support?"
}