Place the reference image over your main video at the first spoken mention, keep the explanation audible beneath it, then end the image layer when the explanation returns to the main shot. In CapCut online, keep the reference on its own image layer so the rest of your narration stays in place.
The image should answer the viewer's immediate question: “What are you pointing at?” A full-frame diagram works well when a small detail matters. Keep that detail and its source credit readable throughout the insert.
Choose the exact image your words refer to
Start with your main video and the image you own or have permission to use. For a chart, photograph or document, keep its identifying information with the file. Include any required credit in the visible image before importing it, so the credit enters and leaves with the reference.
Our example uses an original illustrated desk-light explainer with synthesized narration. The reference is a separate original diagram of a lamp hinge, with an amber circle and the credit “Source: Original demo.” These are teaching illustrations rather than photographs of a real product.
Choose the view that matches the sentence. If you say “the amber circle marks the hinge,” show that circle and hinge large enough to recognize. A decorative picture of a similar lamp would leave the reference unresolved. A crowded diagram needs a simpler version or a closer view that preserves its meaning and credit.
Mark the mention and the return before arranging the image
Play the explanation and note two points: where the image is first mentioned and where the speaker resumes the main scene. Write down the exact words as well as the time. Words help you find the same moment after an earlier edit shifts the timing.
For the illustrated example, the reference passage is: “Look at this diagram. The amber circle marks the hinge. This is where the lamp head pivots.” The return sentence begins: “Back at the desk, the lamp directs light onto the notebook.”
Use the mention as the image's entry point. Keep it on screen while the narration describes its relevant detail. End it before the return sentence needs the main picture again. The diagram passage and the reading task determine the hold; a fixed duration copied from another video may cut off the explanation.
Add image to video projects in CapCut online
Open a video editing project in CapCut online on your Mac. In Media, select Upload to add the video and reference image. CapCut's online MP4 editor supports importing video, audio and image files into the same editing project.
Put the main video on the main track from the beginning. If the image lands after the video during import, drag it onto a separate track above the main clip. Move that image clip horizontally until its left edge aligns with the spoken mention. Keep the main clip in place beneath it so its audio and later frames remain continuous.
In this example, the narration is a separate audio file starting at zero. If your speech is already part of the main video, leave that original sound with the video. You do not need to extract it just to show a reference image.
Select the image and check its framing in the preview. Make it large enough to read the detail under discussion. Preserve the whole diagram when its surrounding labels are needed; use a closer prepared image when only one area matters. Avoid covering the detail with an extra title.
Trim the image's right edge to the return point. A hard cut gives a clear reference change. Once the image layer ends, the underlying main clip becomes visible again. If the insert appears only after the main video finishes, move it above the main clip rather than leaving it appended on the same track.
Check the hold and the return as one continuous passage
Play from just before the cue through the return sentence. Watch for three things: the correct image appears when it is named, the viewer has time to find the detail, and the main scene returns while the explanation continues.
In the example, the amber circle, hinge and source line are the details to inspect. Keep them readable at the size the video will actually be watched. If a phone-sized preview makes the diagram too small, prepare a simpler reference image and use the same entry and exit points. Enlarging the entire layout until the source line disappears solves the wrong problem.
If the image enters late, move only its left boundary or position back to the cue and replay the sentence. If it disappears mid-explanation, extend its right edge. If it stays over the return sentence, shorten it. After each adjustment, check the passage on both sides of the cut so one repair does not hide the next scene.
A silent review helps you judge whether the hinge is easy to locate. Then listen with the picture playing to judge whether the image and words arrive together. These checks answer different questions.
Export a short version and check the saved file
Choose Export → Download in CapCut Web. Give the file a recognizable version name and review the resolution, frame rate and format in Export settings. Our illustrated test uses a landscape frame and a 720p MP4 export at 30fps. Match your own delivery settings to the source and the intended viewing size.
Open the downloaded video and replay the entire reference passage. Check the opening scene, the reference image and the returned scene, rather than relying on the export thumbnail. Keep a small timing record so another editor can repeat the check.
The input narration places the cue at 4.00 seconds and the return sentence at 12.50 seconds. The saved file changes picture at approximately 4.07 and 12.57 seconds; the complete file runs for about 18 seconds. These are measurements from this example. For your own explanation, keep the spoken cue and return phrase in the record alongside the measured times. If a downloaded file shows different content or timing, return to the named project and export the corrected version.
For a separate illustrative shot that needs newly generated motion, start with the CapCut AI video generator. Keep a quoted diagram static when viewers need to inspect its exact details. The AI video creation guides can help you choose the next production task after this insert is finished.
Frequently asked questions
Should the reference image replace the main video clip?
Put it on a separate layer above the main video. The underlying clip can keep playing, including its speech, while the image is visible. Ending the image layer reveals the main video again.
How long should the image stay on screen?
Keep it visible for the sentence that explains the relevant detail and enough time to find that detail. Use the spoken return to the main scene as the exit cue, then review the passage at normal playback speed.
How do I keep the source credit visible?
Prepare the reference with its required credit inside the image, away from the edges. After placing it, check the complete credit at the final viewing size and again in the downloaded video.
What if the image covers the next part of the explanation?
Shorten the image clip's right edge to the return point. Replay from before the cue through the next sentence, checking that the reference remains visible for its own explanation and the next scene appears when needed.