How to Make an EPIC Dunk Video

Making an EPIC dunk video is mostly a photo-selection and template-selection task. The arena motion is already built. You choose the people, the generator prepares full-body references automatically, and the fixed scene handles the drive, jump and finish.

This guide covers the complete EPIC Dunk AI workflow, including the optional second-person route and the mistakes that most often weaken a slam dunk AI video.

1. Choose the right template

Start by deciding how many people should appear in the featured action.

• Solo Arena Dunk: one main player, one photo, a 15 second drive and one-hand finish.
• Duo Arena Finish: a lead player plus a featured defender, a 12 second contested finish and support for an optional second photo.

The duo template always shows an optional second-photo slot. Leave it empty when you want only one uploaded person in that scene; the prepared defender remains.

2. Pick a clean main photo

The main player photo is required. Use one recent image with the person facing the camera and their face clearly visible. Even lighting helps. So does a simple background, although the full-body preparation step will clean that up.

Avoid:

• Sunglasses or a mask covering the eyes
• A face turned far to one side
• Heavy beauty filters that erase skin detail
• Screenshots with sharpening or compression artefacts
• Several people where it is unclear who the image is about

A plain portrait is enough. You do not need to pose like a basketball player or copy a dunk stance. The fixed scene supplies the movement.

3. Add the optional second person correctly

Add one clear photo of the second participant when you want a featured duo. This person becomes part of the scene. Use the same photo-quality rules as the main image.

If you do not have a second photo, leave the optional slot empty. The scene keeps its prepared defender and the main player still appears in the lead role.

Do not upload photos of people who have not agreed to be included. A basketball edit is still a public-facing use of someone's image.

4. Keep full-body preparation enabled

During preparation, EPIC Dunk AI automatically creates one clean studio portrait per selected person. Each portrait shows the person from head to shoes on a solid background.

This step helps the video scene keep body proportions and clothing consistent. It is especially useful in a basketball dunk AI video, where the subject crosses the frame quickly and leaves the ground.

A cleaner reference usually produces a more stable final clip, so this automatic preparation is part of every render.

5. Choose format and resolution

Both templates are built as vertical 9:16 clips. Keep Auto or 9:16 when publishing to TikTok, Instagram Reels or YouTube Shorts. Other frames are available, but a wider crop can reduce the impact of the jump because the player occupies less of the screen.

Resolution changes the credit cost:

The generator shows the exact cost before you submit. There is no free generation tier. Credits are charged when the paid task starts.

6. Generate and review the clip

Submit the task after checking the photos, template and resolution. The status panel reports progress while the full-body references and video are prepared. When the clip is ready, watch it before downloading.

Look for four things:

1. Does the main player still resemble the uploaded person?
2. Does the body stay coherent through the jump?
3. Does the finish land at the rim at the expected time?
4. If you added a second photo, does that person read clearly?

7. Download and publish

Download the finished MP4 and share it in the format that matches the destination. Caption the post honestly as an AI-generated basketball edit. Use the Basketball AI Trend framing when it fits, and avoid presenting the clip as real game footage.

If the video is for a friend, ask before tagging them. If it includes another person, make sure they are comfortable with the post.

Common mistakes to avoid

Using a portrait where the face is too small. The person should occupy a useful part of the frame.

Adding a weak second photo. Leave the optional slot empty or upload a clearer image.

Choosing a wide format for a vertical feed. The source clip is composed for 9:16. A wide canvas weakens the framing.

Blaming the template for an unclear source image. The fixed scene cannot recover detail that was never in the photo.

Forgetting the optional nature of the duo slot. You can still make a solo result inside the duo template by disabling the second person.

Try the complete workflow

The fastest way to learn the process is to run it once. Choose a clear photo and start with 480P or 720P; full-body preparation runs automatically. When you are ready, make an EPIC dunk video with the generator and compare the solo and duo scenes.