Tactiq achieves real-time transcription by using content scripts (basically scripts injected directly into the meeting webpage) along with background processes that handle audio or caption data. It mostly leverages the built-in live captioning features provided by platforms like Google Meet and Zoom, since these captions are already accurate and identify speakers clearly.