Hear a recording on my Mac
Turn one recording in the person's Trillion files into timed words on their own Mac, with the speech model it runs, instead of a speech service Trillion pays for
Runs on the person's Mac: reads the recording from their files (file, by name or identity) or from the Mac (path) and hears it with the first hearing program the Mac has (computer-sources.json hearing rows, such as mlx-whisper). Answers the transcript { text, segments with word timings, language, heard_whole }. Trillion asks it first when one of the person's Macs is awake (src/placement.json), so recordings from a phone or a browser are heard for free.
What it asks for
- file (Optional, words): The recording in the person's Trillion, by name or identity
- path (Optional, words): The recording on the Mac
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 turn one recording in the person's Trillion files into timed words on their own Mac, with the speech model it runs, instead of a speech service Trillion pays for”
What your AI calls: run_tool {"tool_id":"tool:hear-a-recording-on-this-mac"}