Awesome Record (voice + sound)

Use to open an audio recording widget when the user wants to capture microphone input directly inside the ChatGPT conversation. This tool is appropriate when the user asks to record audio, make a voice memo, capture a spoken note, or open a simple recorder UI that they can operate themselves. The tool does not require any input arguments because the recording interaction happens inside the widget after it is displayed. Once opened, the widget lets the user start recording, pause and resume, finish the take, and then either keep playback local or upload the completed recording directly from the browser to configured storage so it can be shared back into the conversation. The tool is intentionally limited to presenting that recorder interface. It does not transcribe the captured speech, perform server-side editing, or attach audio bytes directly to the tool response. Do not call this tool for requests that can be answered in plain text, for users who explicitly do not want a widget, or for workflows that require automatic audio processing rather than an interactive recorder. Also avoid using it when the user asks for a different media action such as playing audio, editing existing audio, or generating synthetic speech. Use it specifically when the best next step is to display a microphone recording widget and let the user control recording themselves. The output and metadata only include a timestamp sentinel confirming that the widget invocation completed; the actual audio data remains in the browser until the user chooses to finalize and share the recording from the widget.

record

Use to provide the record widget with a presigned Cloudflare R2 upload URL so the widget can upload a completed audio recording directly from the browser. This tool is for app use only and should not be invoked directly by the model in normal conversation.

record-sign-upload-url