Here is a question most teams never ask before adding a translation or note-taking tool to their calls: where does the transcript go? For the majority of meeting-bot apps, the answer is "our cloud" — your spoken words, captured, transcribed, and stored on infrastructure you do not control.
What you actually agreed to
When a bot joins to translate or take notes, the meeting audio is streamed off to a third party, turned into text, and retained — often to train models, power search, or simply because deleting data is nobody's priority. For a casual catch-up, fine. For deal terms, salaries, patient details, incident post-mortems, or unreleased product plans, that is a standing copy of your most sensitive conversations sitting somewhere else.
The bot in the room is the tell
If a tool needs to join your meeting as a participant, it almost certainly routes the conversation through its servers to do its job. The visible bot is the honest signal of an invisible data flow. It is also why some people quietly stop talking when the recorder shows up.
A different default: on-device
Translation does not have to work this way. Retalk runs as a virtual microphone and speaker on your Mac — no bot joins the call — and keeps transcripts on the desktop app. The web dashboard holds billing metadata and recent calls, not your conversation audio or text. The translation itself streams in real time across 70+ languages, so you give up nothing on capability to keep your words local.
What to ask any meeting tool
- Does it join the call as a participant?
- Where is the audio processed, and where is the transcript stored?
- How long is it retained, and is it used for training?
- Can you keep the transcript on your own device?
The answers separate a convenience from a liability. If on-device matters to you, try Retalk for Mac.
