get_recording
ChatGPTRead one recording in TOON format; find ids via
list_recordings. Verify candidates with maxLength: 2500; titles alone can be ambiguous. Fully read only relevant recordings. Time ranges return whole overlapping turns, or empty text if none overlap. Keep the time range and timestamp setting unchanged between character chunks. renderedText includes manual edits and inline speaker names. artifacts lists existing summaries/translations; when one answers the request, read it with get_recording_artifact instead of the full text. language is the source language; renderedLanguage is the returned text's language. renderedText is null unless status: done; statusMessage explains processing/failure. renderedTextLength: selected text character count (0 if null).





