LiveFaceSwap.co

Guide

How to swap faces in real time in your browser

Choose a reference, set up your camera, go live and read the timer: what to expect from latency, output quality, consent and credit settlement.

Updated 2026-11-07

What you need before you start

You need a computer with a webcam, a current desktop browser (Chrome, Edge or Safari), a stable internet connection, and an account with at least a few credits. LiveFaceSwap.co runs each session on a cloud model, so there is no graphics card to check and no software to install. New accounts get 10 free credits, which is enough for a short first session.

Step 1: choose a reference image you are allowed to use

The reference image tells the model what to put on your face. In Face Swap mode it is a face; in Restyle mode it is a style reference; in Try-On mode it is a clothing image. The studio offers six quick picks per mode plus a gallery you can open when none of the six fit.

  • Use a front-facing photo with the eyes, nose and mouth clearly visible, even lighting, and no sunglasses, mask or hair covering the face.
  • Use a face you own or have explicit permission to use. Do not use celebrities, public figures or anyone who has not agreed to appear in generated video.
  • Higher resolution helps, but a sharp, well-lit face matters more than a large background. A clean head-and-shoulders crop usually tracks better than a wide shot.

Step 2: set the camera and the light

Open the live studio and press Open camera, then allow camera access when the browser asks. Sit facing a window or a soft lamp so your whole face is lit evenly; a bright window behind you turns your face into a silhouette and the swap will struggle. Keep your head inside the frame with a little space around it, and keep the camera still.

Close other apps and tabs that hold the camera. Browsers normally allow one application to use a webcam at a time, so if the preview stays black while your camera light is on, a call app or another tab may still own the device.

Step 3: confirm rights, consent and disclosure

Before it will start, the studio asks you to confirm that you own the reference image or have the person's permission to use it, and that you will not use the result to impersonate or deceive anyone; the acceptable-use policy linked beside the checkbox applies to every session. That confirmation is not a formality. If the face is not your own, you need the person's permission, and anyone watching should know the face on screen is AI-generated. Several platforms and jurisdictions require synthetic video to be labelled, and using someone else's face without consent can be illegal even when it is a joke.

Step 4: start the session and read the LIVE timer

Press the start button in the output stage. The browser sends your camera frames to a cloud session running LiveFaceSwap V1, and the swapped video appears beside your camera. Starting a session reserves credits for up to 900 seconds of live time, and the LIVE timer in the output stage shows how long it has been running.

The output is generated in the cloud, so it trails your real movement by a short delay that depends on the network. A wired connection or a strong 5 GHz Wi-Fi signal keeps the delay and the stutter down. Every live second is billed at 0.6 credits, and a session stops at the 900-second cap; if you need longer, start a second session when the first one ends.

Step 5: stop the session and settle the credits

Press Stop when you are done. The provider's own meter reports the seconds consumed, the platform bills 0.6 credits per live second rounded up to the next whole credit, and the unused part of the reservation is returned to your balance immediately.

Do not just close the tab. A closed tab can leave the session running until it expires and is settled by the cleanup sweep, which is a bad way to discover how billing works. Pressing Stop is the clean end of a session, and it is the only way to get the unused reservation back right away.

What the output is, and what it is not

Live face swap is a generated video preview: it follows your expression and head movement, but it is not a pixel-perfect copy of the reference and it can show soft edges, artefacts around the hairline, or mismatched lighting when you move quickly. That is normal for a model running in real time, and different from an offline video render that can spend minutes or hours on a single clip.

The studio does not record your camera stream. The session processes live frames, and it stops when you press Stop.

FAQ

Does real-time face swap need a GPU?
No. The model runs in a cloud session, so a normal laptop with a working webcam is enough. A graphics card on your machine would not be used by the browser studio.
Why does the swapped video trail my movement?
Frames make a round trip: browser to the cloud model and back. Network delay, upload bandwidth and a weak Wi-Fi signal all add to it. Move closer to the router or switch to a wired connection before assuming the model is slow.
How many credits does a 10-minute session use?
Ten minutes is 600 live seconds, so 600 x 0.6 = 360 credits, billed from the provider's meter and rounded up to whole credits. A 1,000-credit pack is about 28 minutes at that rate.

Try the live studio or see pricing and credit packs.