Renames a recording and, when you send file, replaces its audio. name is required on every call, even when you are only changing something else, and the text-to-speech fields are replaced wholesale: tts_text, tts_ssml, tts_voice, tts_provider, and tts_editor_state are cleared whenever you omit them, and tts_speed and tts_instructions are always cleared. Sending new audio deletes the previous file first, converts the new one to WAV, and — unless you also send tts_ssml — transcribes it and stores the transcript as the recording's text. A recording a call flow owns cannot be edited here at all, and a recording a call flow uses cannot have its audio replaced; both are refused with 409. Requires an API key with the Admin role.