Image generation and audio transcription

Agent image generation and audio transcription use AI tools. They are separate from Portal voice input and meeting features.

Before a media task

1. Connect the required OpenAI credential.
2. Verify that the corresponding tool is ready. A saved credential alone is insufficient.
3. Specify the input, output, budget, and retention requirements.
4. Run a small test and check the actual output format.

Never put secrets in Chat. Confirm rights to other people's voices and materials. Do not assume results automatically become Tasks; see files and artifacts.

For Portal voice input, follow Chat Rooms. For recording and meeting summaries, follow meeting minutes.

If processing fails, check the connection, tool readiness, format, and budget. Check whether processing already occurred before retrying. Disconnect unused credentials and remove retained inputs and outputs according to policy.