YouTube thumbnail guide
How to Turn a Portrait Photo Into a YouTube Thumbnail
A vertical portrait does not need to lose the top of a head, shoulders, or clothing to fit YouTube. Expand the scene to a wide canvas, protect the person, and design the new space around the thumbnail message.
Quick answer
To turn a portrait photo into a YouTube thumbnail, place the full portrait inside a 16:9 canvas, move the subject to one side, and use AI outpainting to generate background in the empty area. Add short, high-contrast text only after expansion, then export a large JPG or PNG that accurately represents the video.
What changes when a portrait becomes a YouTube thumbnail?
A portrait photo is taller than it is wide. A standard YouTube thumbnail is usually a wide 16:9 composition, so simply enlarging the portrait cannot fill the frame. You must either crop the person, add plain padding, or create new visual content around the original.
AI outpainting uses the third approach. It keeps the source photo at its natural proportions and generates a continuation of the background in the new canvas area. That extra width can also become intentional negative space for a title, number, arrow, or other small design element.
Plan the portrait and text before expanding
This project-owned portrait illustrates the source you want to protect. Decide where the face and headline will sit before asking AI to fill the wide canvas.
Headline space
Put the message opposite the face
Which method should you use for a vertical photo?
| Method | Keeps the portrait | Fills 16:9 | Best use |
|---|---|---|---|
| AI outpainting | Yes, if the source is protected | With generated surroundings | Natural, full-frame thumbnail layouts |
| Crop | Only the visible center | Yes | Close-ups with disposable outer areas |
| Solid padding | Yes | With a flat background | Simple graphic or branded layouts |
| Stretch | No; proportions change | Technically | Avoid for portraits |
How to turn a portrait photo into a YouTube thumbnail in five steps
Start with the original portrait
Use the highest-resolution original you have, not a screenshot from a social app. Confirm that the face, hair, hands, clothing, and any product in the photo are already sharp. Expansion can create surroundings, but it cannot reliably recover detail missing from the subject.
Open a 16:9 canvas
Choose a 16:9 preset and keep the entire portrait inside it. Do not stretch the photo to fill the width. The blank area around the portrait is intentional: it gives the outpainting model room to continue the scene while preserving the person.
Place the person opposite the future headline
Move the portrait left when text will sit on the right, or right when text will sit on the left. Keep the face away from the outer edge and leave uncluttered negative space for a short title. Make this composition decision before generating the background.
Generate only the missing background
Use AI outpainting to continue the existing room, studio backdrop, sky, wall, or other environment. If prompts are available, describe the background, lighting, color, and depth rather than asking the model to recreate the person. Regenerate any result that changes the subject or adds distracting objects.
Add text and export the thumbnail
Add a short headline in a design editor after the image is expanded. Use strong contrast and check the design at a small preview size. Export a large 16:9 JPG or PNG, verify the file against YouTube's current upload guidance, and confirm that the image accurately represents the video.
YouTube thumbnail export settings
YouTube's current custom thumbnail guidance recommends a 3840 × 2160 image, a minimum width of 640 pixels, JPG, GIF, or PNG, and a 16:9 ratio for the format most used in YouTube players and previews. Its listed file-size limits differ between mobile and desktop, so check the source at upload time rather than relying on an old template.
- ✓ Aspect ratio: 16:9 for a standard long-form video thumbnail
- ✓ Working size: 3840 × 2160 according to current YouTube guidance
- ✓ Accepted formats in YouTube's guidance: JPG, GIF, or PNG
- ✓ Design check: the face and headline still read at a small preview size
- ✓ Accuracy check: the image honestly represents the video's content
YouTube also prohibits misleading or otherwise disallowed thumbnail imagery. Read the current thumbnail policy before publishing.
How to keep the expanded background believable
Protect the entire person: keep hair, shoulders, arms, and clothing away from the generation boundary. The closer a body part is to that boundary, the easier it is for the model to alter it.
Continue existing visual cues: match the source light direction, background blur, horizon, wall lines, and color temperature. A simple continuation is usually more convincing than a new scene.
Leave useful empty space: do not fill every generated area with objects. Calm negative space makes the portrait and headline easier to understand at thumbnail size.
Inspect before adding text: zoom in on the old photo boundary and check for repeated patterns, broken edges, floating objects, or changes to the person. Fix the image before designing on top of it.
Related image expansion guides
For a deeper explanation of canvas math and wide framing, read how to extend an image to 16:9 without cropping. If you are choosing an editor first, compare the best AI image extenders.
Frequently asked questions
Can I make a YouTube thumbnail from a vertical photo without cropping?
Yes. Put the complete vertical photo inside a 16:9 canvas and generate background in the empty area. This preserves the original person instead of trimming the head, shoulders, or clothing. You can then position the portrait on one side and reserve the other side for a concise headline.
What size should a YouTube thumbnail be?
YouTube currently recommends a 3840 × 2160 image for a custom long-form video thumbnail, with a minimum width of 640 pixels, and says 16:9 is the most-used aspect ratio in its players and previews. Check YouTube Help before exporting because upload recommendations and file-size limits can change.
Should I stretch a portrait photo to 16:9?
No. Stretching changes the person's proportions and makes faces and objects look unnaturally wide. Keep the original portrait at its natural proportions, enlarge the canvas to 16:9, and fill the added area with a continued background instead.
Where should the face go in a YouTube thumbnail?
Place the face on the side opposite your text, with comfortable space around the head and eyes. There is no universal position for every video, so test both left- and right-aligned layouts. The strongest option is the one that remains legible and balanced at a small preview size.
Can AI change the person's face while extending the photo?
It can, especially when the face touches the generation boundary or the tool redraws too much of the original. Keep the complete person inside the protected source area, ask the model to generate only the background, and reject outputs with altered facial features, hair, hands, clothing, or logos.
Expand your portrait for a YouTube thumbnail
Upload the portrait, choose 16:9, position the person for your layout, and let AI Image Extender create the background space you need without cropping the subject.