Keep a recording as a voice memo and hand it to your AIs
Make a recording already in this person's files into a voice memo: keep its words beside it as a transcript file and hand it to their AIs to act on, the way the Record screen does
Give identity, the recording's identity in the person's files (audio or video). Give words and seconds when the words were already heard (a live transcript); without words its own transcript is made (transcribe-file) and its length read from it. Give format_id and about when it was read from a format's script, so the transcript starts with the script and the AI turns it into that format. Keeps "<name> transcript.txt" beside the recording and asks the person's AIs to do something useful with it; the same recording again is handed on once. Returns audio_file, transcript_file, transcript_identity, words, asked_id, repeated and said; a recording in which nothing was said keeps its empty transcript and answers nothing_said true, handing nothing on.
What it asks for
- identity (Needed, words): The recording, from find_file or list_files
- words (Optional, words): What was said, when it was already heard
- seconds (Optional, number): How long the recording runs; needed with words
- format_id (Optional, words): The format whose script was read, from content_formats
- about (Optional, words): The thought the script was made with
Use it from your AI
Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this tool by what it does and runs it.
Say to your AI: “Use Trillion to make a recording already in this person's files into a voice memo: keep its words beside it as a transcript file and hand it to their AIs to act on, the way the Record screen does”
What your AI calls: run_tool {"tool_id":"tool:keep-a-voice-memo"}