Notion has officially launched Speaker Attribution v1.0 for AI Meeting Notes, a major update that enables the system to identify and label who is talking during 1:1 calls and certain video conferences.

This update addresses one of the most significant gaps in Notion’s meeting suite: the “wall of text” transcript. By utilising dual-stream audio capture—processing both your microphone and your computer’s system audio—the AI can now distinguish between participants. In the initial v1.0 rollout, this feature is optimised for 1:1 meetings, where it can cleanly separate the “local” and “remote” voices.

As the transcription begins, you will see that speaker labels appear next to each new line of text. While currently focused on 1:1s and specific conferencing platforms, this foundation sets the stage for more complex multi-speaker identification in future updates.