Skip to content

Recording a lecture

Record audio from a note and follow the transcript as it arrives.

Recordings belong to a note. Open a note, go to its Recordings, and start recording. WorkRoad records your microphone and can mix in tab or system audio alongside it.

Recording uses AI work and consumes credits.

While you record

Audio is captured in roughly 60-second, self-contained batches. WorkRoad sends each batch to the server as soon as it closes, and the transcript grows while you continue recording.

Where supported, the browser's own speech recognition shows a live preview of what you say into the microphone. That preview does not cover tab or system audio. The server transcript replaces it when the current batch arrives.

Drop a marker when you want to flag a moment. Pausing closes the current batch, and resuming starts a new one. This makes a pause durable. A silence detector for each audio source can also end a batch.

A recording session can last up to four hours.

Reloading and navigating

The recorder widget keeps running while you move around the app. If you reload, earlier transcripts return and at most the batch still in flight is lost. You must share tab audio again after a reload because the browser does not preserve that permission.

Review the recording

After recording, the transcript stays anchored to the audio timeline. Click a transcript line to seek to that point. During playback, the line currently being spoken is highlighted.

You can also start a generation from your sources. It runs as a durable background job, shows live progress, and is metered in credits.