How to record your screen and your webcam together on Windows

All articles

A screen recording with your face in the corner is easier to watch than one without. People follow a walkthrough better when they can see who is talking. The catch is the framing: you pick a corner and a size before you start, and if it covers the one button you were pointing at, the usual answer is to record it all again.

AgentEyes, our screen and audio recorder for Windows, treats the framing as a decision you can change after the recording is made. This is how.

Set it up once, as a preset

AgentEyes records from named presets, in the spirit of OBS scenes. A preset says which monitor or region to capture, whether it records video, audio or a single screenshot, how the audio is handled, and whether a webcam is filmed alongside the screen.

  • The camera. Name a webcam and a frame rate in the preset.
  • The framing. Choose a circle or a rectangle, which corner it sits in, how big it is, and which part of the camera frame the circle takes. You choose it in the preset editor against a live picture from the camera, so you see what you will get.
  • The audio. Microphone only is the default, because it captures the room correctly timed. Noise suppression and voice leveling are applied to the microphone unless you turn them off. Mixing in the system audio is there when you want it.

Then pick the preset and record. A recording indicator is on the whole time it is capturing, and one action stops everything, from the window, the indicator or the tray.

The camera is always kept whole

This is the part that saves the second take. Whatever framing you picked, AgentEyes records the webcam to its own camera.mp4 beside the screen video, in full. The circle in the corner is drawn from it afterwards.

When you stop, the work carries on without you: finishing the audio mix, rendering the camera into the video, making the thumbnail, and transcribing and titling the recording all run as one sequence after the stop, so the app answers straight away. Transcribing and titling use DevThrottle's hosted AI, so those two steps need you signed in. If that sequence is interrupted - you close the app, the machine restarts - it picks up at the step it reached the next time AgentEyes runs.

Change the framing afterwards

Watched it back and your face covers the menu you were explaining? Change the framing, then render the video again from the recording folder:

agenteyes compose <recording-folder>
  • The composed video is written to recording.mp4 at the new framing.
  • The screen-only cut is kept as recording.screen.mp4, so you also have a version with no camera.
  • The camera file is never changed, so you can run compose as many times as it takes.

Compose runs on your own machine. A recording with no camera track is reported and skipped.

From a script instead

The same engine is a command-line tool. agenteyes video takes --camera "NAME" to film a webcam beside the screen, and --camera-fps if you want a rate other than the default. While a session runs, S takes a screenshot and Q stops.

Where it lands

Each recording is its own folder under %USERPROFILE%\Videos\AgentEyes\, holding the screen video, the camera track, any screenshots you took along the way and, once the sequence has run while you are signed in, the transcript. AgentEyes is its own download, and you do not need a DevThrottle account to record.

Presets, the camera and the framing are covered on the Recording page of the docs. Every command, including compose, is in the AgentEyes CLI reference.

Talk to usBook a 20-minute call