Set up OBS face swap with your live camera and a character reference. Preview the face change in the studio below, then capture its browser window in OBS and place it over your tutorial window. Record the combined screen, character and microphone in OBS.
Sign up to try it with free credits.
Checking browser capabilities…
Checking live preview support…
Make a short drawing or slide tutorial: start with a square, add a circle and move it beside the square. Your document is the main picture; the character explains the action from a smaller corner window. The useful result lets someone follow the pointer, see both shapes and hear each instruction without the facecam covering the work.
Use the complete Live Face Swap studio above to create your character picture. In this OBS face swap walkthrough, OBS records the tutorial window, facecam and voice together, so you can review the explanation as a viewer would.
The walkthrough below uses OBS Studio on Windows and two separate application windows. This website supplies the live AI picture; OBS captures and combines it with your document. It does not install an OBS plugin, publish a dedicated Browser Source feed or create a system camera. Capture behavior depends on your setup; the complete OBS recording workflow has not been tested here.
| Source | What it contributes | Check it with your own input |
|---|---|---|
| Tutorial WindowWindow Capture | The drawing or slide app, large enough to read its controls and see the pointer. | Move an object in that app. The same change should appear in the OBS composition. |
| AI FacecamWindow Capture | The studio window, cropped to the AI output and placed above the document source. | Your current head movement reaches the corner picture while the document stays visible behind it. |
| VoiceAudio Input Capture | One selected microphone for the spoken explanation in this scene. | The OBS meter responds to speech, and the saved test contains one clear copy of the line. |
An OBS face swap tutorial needs separate sources for the work, the character picture and your voice. Name them clearly so a wrong window or duplicate microphone is easier to locate. Use a scene named Tutorial for the example below.
Prepare the document, capture windows and voice with AI off. In your own drawing or slide app, put a square near the middle of a blank page. Leave room beside it for a circle. Read the walkthrough first so you can keep the studio at its tool position once you crop the facecam.
If the browser cannot open your intended camera, use the Live Face Swap PC guide to check Windows permissions and device selection first.
Create the Tutorial scene. In Sources, choose + → Window Capture, create Tutorial Window and select your drawing or slide app. Enable Capture Cursor in this source's properties. Keep the studio in a separate window so selecting a tutorial tab does not replace the captured AI picture. Check the chosen window in OBS before adding the facecam.
Point at the square, move it slightly and undo that movement. Confirm that OBS shows this document and its current pointer action. If the image is blank or the wrong app appears, fix the window selection now. Adding a character cannot repair the document source.
Size the document source for your intended recording canvas. Keep the controls used to add a circle and the final positions of both shapes in view. In OBS Settings → Output, note the Recording Path and format so you know where the test will be saved.
Choose a lower corner with no required control or drawing in it for the future facecam. Move the pointer through the full action, including the circle's final position. If the important details are already too small, enlarge the document view or simplify the workspace before starting AI.
For this scene-specific example, disable any global Mic/Aux input using the same microphone in Settings → Audio. Add an Audio Input Capture source named Voice to Tutorial and choose your intended device. Speak a sentence and check its OBS meter. Mute unrelated desktop or captured application sound for this silent-document exercise.
Make a brief OBS recording with AI off. Speak the sentence while moving the pointer around the square, then stop and open the file in the Recording Path you noted. Check that the pointer moves and the whole sentence is audible once. The studio's Record microphone option controls studio clips; OBS captures this narration through its own selected input.
Add the OBS face swap picture only after the document and microphone checks are ready. Keep this studio selected in its own browser window while you work in the other application. Avoid scrolling, changing browser zoom or resizing that window after the crop is set.
Add another Window Capture source named AI Facecam and select the browser window containing this studio. At the tool, use Transform → Edit Transform in OBS to crop around the output area, excluding the browser tabs, camera-input panel and settings. Use the idle example to locate the area while AI is off.
The idle example only helps you arrange the capture. It is not evidence that your camera or live OBS face swap is working. Keep the studio window at this size and scroll position; revisit the crop after any layout change.
In the studio, select Face Swap and choose a reference or upload your own. Open the camera and sign in. Saved preferences may reopen a different effect, so check the active mode. Turn off Auto record if you want only the combined OBS recording. Review the model's displayed rate and your balance, then start AI.
Wait for your own output and make a small head movement. Confirm the same movement in OBS, then focus the tutorial app and check that the facecam keeps updating. If the picture stops during this switch, inspect the capture and browser state before recording. Choose End while resolving a longer setup problem.
Put AI Facecam above Tutorial Window in the Sources list. Resize and move it into the corner you checked earlier, preserving the picture's proportions. Inspect all edges for the original camera panel or controls, then lock the positioned sources to avoid dragging them accidentally while working.
Repeat the pointer path for adding and moving the circle. An OBS face swap layout is ready for the test when the pointer, required controls and both object positions remain clear. If the overlay hides a step, change its position or size before starting the tutorial recording.
After placing your own moving character over the document, judge both at the size of the finished recording. The face should be readable without taking space from the drawing controls or the completed shapes. The room around the character remains part of the captured picture, so allow space for the whole rectangular facecam.
Check the selected model's displayed rate before starting. Use Pricing to compare available credits, and choose End while you work on a longer setup change.
Use the three passes below for a short OBS face swap recording. These are suggested actions and lines to perform in your own app, not a transcript or measured result. Say what you are about to do, leave a small pause, then make the change so the viewer can follow the pointer.
Put Tutorial in the actual OBS output. If Studio Mode is enabled, send the prepared scene to Program with Transition. Choose Start Recording in OBS, return to the document and point at the existing square. Begin speaking after the recording starts.
Practice line
The square is already on the page. I'll add a circle beside it.
The saved opening should identify the square clearly and include the entire first sentence. The viewer should see the same page you are describing, with your current character in the corner.
Say the first instruction, pause briefly and add the circle using your app. Then explain the move, pause and drag it beside the square. Keep the path and the destination away from the facecam. Leave the new position visible before continuing.
Practice line
Add the circle here. Now move it beside the square, leaving a small gap.
Watch whether each instruction can be connected to one visible action. If the pointer or destination disappears behind the character, correct the composition. If the voice refers to a change before the viewer can locate it, repeat the delivery more deliberately.
Show both shapes and the gap while you say the closing line. Leave a short pause, then choose Stop Recording in OBS and wait for it to finish. Choose End in the studio to stop AI. Open the completed file from your OBS Recording Path.
Practice line
Now both shapes are visible, with a small gap between them.
The closing words and final arrangement should be present before recording ends. Stopping OBS does not end the AI session, and ending AI first can replace the moving character while OBS is still recording. Inspect the finished file before using this setup for the longer tutorial.
Open the file saved by OBS and watch the whole tutorial with sound. A good studio preview checks the character alone; this file also checks the document, pointer, composition and narration. Keep the take when all three passes communicate clearly, or use the first failed check below to choose one adjustment.
Check narration against the pointer action as well as the moving face. The AI picture travels separately from the document capture and microphone. An audio delay affects the whole voice track: a change that improves the face can make an instruction arrive after the click it explains. If you adjust timing, repeat the same action and judge both relationships in the saved file; one fixed adjustment may not suit the whole take.
Check Capture Cursor on Tutorial Window.
Stop the recording and choose End in the studio. Open Properties for the Tutorial Window source in OBS and enable Capture Cursor. Return to the drawing or slide app and move the pointer across the square. With AI off, make a short recording and check the pointer in the saved file before repeating the full explanation.
Change the composition, then repeat the same action.
Move or reduce AI Facecam while keeping the character recognizable. Check the pointer's entire path, not just its starting position. For another tutorial, choose the corner from that task's controls and result rather than reusing this placement automatically.
Check the source order and the facecam's visibility.
The character source should sit above the document source in the Sources list. Check its eye icon and its position within the canvas. A correct browser picture will still be hidden by a full-size source placed above it.
Recheck the selected window and all crop edges.
Stop the recording while you inspect the source. The capture may include the camera-input panel or may have shifted after the studio window changed size. Correct the crop and make a new test; changing the face reference does not change what part of the browser OBS captures.
Compare the live studio picture with its OBS capture.
If the studio also stopped, check its message, connection and balance. If the studio moves while the facecam does not, inspect the selected capture window and keep the studio tab selected. Confirm movement again while operating the document before starting a replacement take.
Check the Voice source, mute state and duplicate audio paths.
A moving microphone meter is a useful input check, but listen to the saved file as well. Inspect recording audio selection if the file is silent despite a working input. For a doubled voice, check whether the same microphone or monitored playback is entering OBS through a second route.
Recheck the explanation against both moving pictures.
An OBS face swap tutorial has two timing relationships to judge. Compare the instruction with the document change and the moving character over the complete sentence. If one timing adjustment improves one relationship and harms the other, change the delivery or resolve the delay before keeping the take.
Before a replacement take, stop any current OBS recording and choose End in the studio. With AI off, remove the circle you added and return the square to its original position. Keep the same document window, window size and zoom. Recheck the selected capture and pointer, then start AI again. Once your moving character reaches OBS, record from the opening line with only the square on the page.
Keep the setup when the square, new circle and final gap remain visible, the pointer identifies the right place, and the complete explanation is audible once. For your real topic, replace those three checkpoints with the control you introduce, the change you demonstrate and the result the viewer should recognize.
Prepare your OBS face swapFor a public Q&A, continue with Deepfake Avatar Streaming. For a conversation with another person, use Face Swap Video Call. Those guides cover their own output and receiving-end checks.
The studio creates a live AI character picture from your webcam and reference. OBS captures that browser window, crops the output and combines it with the application you demonstrate. Your microphone enters OBS separately. The final tutorial recording is saved by OBS.
This page provides the browser studio and a window-capture workflow. It does not install a plugin or register a camera device. In OBS, capture the existing studio window to obtain its picture. A complete recording still needs to be checked on your own browser, system and OBS setup.
An OBS Browser Source opens a separate browser environment, so the page URL does not connect it to your current signed-in studio or camera session. The walkthrough uses Window Capture on Windows. Keep the studio selected in its own window and demonstrate your task in another window.
This workflow captures a rectangular AI picture, including its surroundings. Cropping removes unwanted edges; it does not automatically isolate a transparent character. Leave space for that rectangle when you choose where it sits over the document.
Choose the narration microphone in OBS and keep one intended route for that device. The scene-specific example uses Audio Input Capture with the duplicate global input disabled. Record a short test and listen to it. Record microphone in the website controls recordings made inside this studio, not the OBS audio mix.
OBS saves the combined document, facecam and voice to its own Recording Path under Settings → Output. The studio clip list contains only clips recorded here. Review the OBS file for this tutorial; a studio-only clip cannot show whether the pointer or document was covered.
Yes. Prepare the OBS sources and microphone with AI off, then check the studio's displayed rate and available credits. Eligible new accounts can try AI with trial credits. Use a short private recording to judge the combined result. Stop the OBS recording and choose End in the studio when finished.
The detailed capture steps here use desktop OBS on Windows. On modern macOS, OBS uses macOS Screen Capture and requires the appropriate capture permissions; source selection differs. A phone browser alone cannot perform this two-window desktop composition. Check the supported capture method on the computer you intend to record with.