Skip to content

Presenter Mode — Anyone On Camera, Over Your Screen ​

October 1, 2026

Record your screen, put a person in the corner explaining it, and ship the file. No camera, no crew, no stitching clips together by hand.

Presenter Mode is the whole path: generate someone on camera, lay that clip over a screen recording, move them between full-frame and a corner box, freeze the screen while they talk, draw on the held frame, and export. Alfred and Ada ship as standard — and now anyone you register can take the same seat, with more than one of them in a single video.

This note covers the feature end to end.

1. Get someone on camera ​

  • Ask for Alfred or Ada by name. A clip that names one of them is generated from their official look. The opening frame is their portrait, and other angles of that same person travel with it. You describe what they say and do — never their face, hair, or clothes, and a different photo is never substituted for them. Put the words after says:; anything after the line, like a smile or a gesture, is performed rather than spoken.
  • Register your own people. Add a colleague's photos to your presenter folder under a name of your choosing, and they are castable by that name exactly like Alfred and Ada — the same consistent face in every clip. One good front-facing photo is enough to start. Full setup is in Presenters.
  • A clip works as well as a portrait. You can register someone from a short video instead of a still. One representative frame is picked automatically, stepping over black lead-in, blinks and motion blur — so two seconds of phone video registers somebody who has no studio photo.
  • Character sheets are cropped for you. A figure inside a flat border with a caption strip underneath is trimmed to the panel automatically, and framed to the clip's aspect ratio with the head kept in shot.
  • Length. Ask for 4 to 15 seconds. If the first attempt is refused, the retry uses 4, 6, or 8 seconds, whichever is closest to what you asked for.
  • Opt out. Say you do not want a presenter. Writing ADA in capitals — the Americans with Disabilities Act — does not cast Ada.

2. Lay them over the screen recording ​

Open Live → Video Studio, pick the presenter clip and the screen recording, and the Studio becomes a two-lane editor.

  • Three layouts. Presenter full screen, Screen + presenter box, and Screen only. Moving between them either animates over 0.4s or cuts.
  • The presenter box. Starts at an eighth of the frame, bottom-right. Pick another corner or drag it anywhere, and set margin, rounded corners, border and shadow. Size runs from 1/16 to 1/4 of the frame.
  • The screen waits for the box. The recording starts playing the moment the box first appears, so a full-screen intro does not eat into it — and the presenter's voice carries straight on over the recording.
  • The preview is the export. What plays in the Studio is what renders, frame for frame.

3. Put two people in one video ​

The Script lane holds a sequence of lines, and each line says who is on camera and what they say. Alfred opens, Ada takes the numbers, your colleague closes — switched on the timeline instead of stitched together afterwards.

Each line carries who is speaking, their line of dialogue (kept with the clip so the script stays readable without opening the video), and the clip that plays. Drag a line's right edge to change how long it holds and the lines after it move with it, so no silent gap opens up. The preview switches presenter exactly where the export does.

4. Control the pacing ​

  • Freeze the screen. Each freeze is 0.25 to 120 seconds (3 seconds when you click Freeze here), and chooses one of:
    • Screen (presenter keeps talking) — the default. The screen holds; they keep speaking.
    • Screen and presenter — both hold.
    • Presenter (screen keeps playing) — only they pause, which can sit before the screen recording starts.
  • A paused presenter does not freeze mid-word. Their speech fades, a silent moment from their own clip dissolves in and holds, then dissolves out when they resume.
  • Mark up a held screen. Arrows, boxes, ellipses, pen strokes, highlights and text draw on a frozen frame. Markup needs the screen paused — it is unavailable when only the presenter pauses, because the screen is still moving.

5. Finish and export ​

  • Screen audio. Mix the recording's own audio under the presenter. Output size and frame rate (15–60) are set in the same panel.
  • Captions. Burn them into the picture, or carry them as a track the viewer can switch off.
  • An ending slide. Close on a card when the demo should not end on a talking face.
  • Export MP4 from the Studio header. The file lands in Generated Files, and the button becomes Download MP4 until you change the timeline again.

Where this shows up ​

You want a product walkthrough with a person in the corner, not a faceless screen recording. You used to generate a talking head, hope the face matched last time, then composite it by hand. Now you ask for Ada, upload the recording, and place her.

You want your own colleague presenting, not Alfred. Three photos in the presenter folder and they are cast like anyone else.

You are mid-demo and need the dashboard to hold while she explains a number. Add a freeze, leave the pause on Screen, and draw on that frame.

Try it ​

  • "Make an 8-second clip of Ada looking into the camera. She says: And I'm Ada. Here's the dashboard."
  • "Make a 6-second clip of Alfred. He says: Here's what changed this quarter."
  • "Put that clip over in/demo.mp4 in Presenter Mode. Start full screen on Ada, then move her to a box in the bottom right when the dashboard appears. Freeze the screen for 4 seconds at the billing page so she can explain it, and draw an arrow at the total."
  • "Put Alfred's intro first, then Ada from 0:06 for the numbers."
  • "Add an ending slide with our logo and burn in the captions."

Heads up ​

  • Registering someone is a consent decision. A presenter is a recognisable person. Add someone only when they have agreed to be generated on camera and to the use you intend — "they work here" is not consent, and a photo you happen to have is not permission. Deleting their files de-registers them immediately.
  • The people you register stay yours. Their photos and clips live only in your presenter folder, and removing them takes that person out of the cast straight away.
  • One presenter per clip. A prompt that names two registered presenters does not pick between them. Make one clip each, then give each its own line on the Script lane.
  • A video with a Script lane takes its presenter from the lane. Videos made before the lane existed are unaffected.
  • Presenter clips and Presenter Mode exports are on Pro and above, same as Video Lab and Video Editor. A presenter clip is one Video Lab generation; a refused first attempt is not charged twice, and a clip that cannot be made is not charged. Exporting the timeline is a Video Editor render like any other rebuild.
  • Other pictures you attach are not used for a presenter's face. Generate the product shot as its own clip, or say you do not want a presenter.
  • Prompts and the portrait go to the video generator the same way as any other Video Lab clip. Runway retains prompts for an unspecified time and does not train on them. The finished file stays in your session.

Built for Alfrada OS.