Let the Camera Follow Each Step

Start with a short product walkthrough

The reference video at the top of this page turns an AI workflow into an animated walkthrough. The camera follows a progress line, moves toward the active step, and shows cards, images, and generated results. Viewers can follow the camera rather than searching the interface for what matters.

The creator shares tips for making this style of animation with Claude. This page turns those tips into a prompt you can copy, with no skill package to install. Choose one short workflow from your product, such as “enter a request → choose an option → view the result.” Use three steps to explain one task.

1. Prepare a screenshot or sample interface content for each step, noting the action and result. Label the screenshots in order. If the interface is not ready, start with text and placeholder cards.
2. Attach the materials in a Claude conversation, send the prompt on this page, and add a brief like the one below. Replace the product, steps, and interface text with your own.

Product: a task-management tool.
Purpose: show how to turn a short request into tasks.
Step 1: enter “Prepare next week's product launch” in an input field.
Step 2: reveal three generated task cards: define the selling points, build the product landing page, and schedule the launch.
Step 3: move the camera toward “Build the product landing page” and expand its owner and due date.
Start with a landscape preview of about 6 seconds showing only these three steps. Do not add other features.
Extend the duration if viewers need more time to read the text.
First explain the implementation approach, how I will open the preview, and how the finished video will be exported. Then outline the shots.
Wait for approval before creating the preview.

3. Approve the implementation approach and shot order, ask Claude to build the preview, and open it as agreed. The first version should be playable or interactive, with visible camera movement. If you only receive a written shot list, ask Claude to build the preview. Check whether the camera moves toward the active step, text stays visible long enough to read, and each action has a visible result. Adjust the order and pauses before adding more decoration.

Image and video generation in the reference are parts of the product being demonstrated. This prompt does not itself provide those services. Use existing result images when building your own walkthrough.

Add depth, glow, and light trails

The creator’s 2.5D approach places flat interfaces and cards on separate depth layers, using camera movement to reveal their spatial relationship. It does not require turning every element into a 3D model.

Get the camera and card movement working first, then add effects one at a time. Continue in the same conversation with:

Keep the three steps, interface text, and total duration. Change only the animation:
Bring the active card slightly closer to the camera and move the other cards into the background.
Ease the camera in and out of each move, keeping it still while text is being read.
When the light on the progress line reaches a node, illuminate the node and then reveal its card.
Give the moving light a short trail. Keep the bloom clear of text.
Let images resolve from coarse pixel blocks into a fully visible, sharp image.

Fix one problem at a time. If text is hard to read, extend the pause or reduce the glow. If the camera feels unstable, reduce rotation. If every card competes for attention, keep upcoming cards still. Then add brief sound effects to key actions, synchronized with the movement.

Use your own workflow and materials

Changing the product means more than replacing its title. Update screenshots, step names, results, and the camera path together. Fill in this brief with your own details:

Product and audience: [product name and intended viewers].
One task to explain: [what viewers should understand by the end].
Step order: [the action and result at each step].
Materials: [which file belongs to each step].
Focus areas: [where the camera should move closer or pause].
Delivery: [aspect ratio, total duration, whether narration is needed, and whether sound effects are needed].
Outline the shots first. After approval, make a short preview before extending it into a full walkthrough.

For a portrait version, rearrange cards, the camera path, and text size rather than cropping the sides off the landscape version. Once the full workflow is approved, ask for export instructions for the actual project. If it only provides an interactive preview, use a screen recorder to capture it first.

Start at 30 FPS and focus on the camera and timing. The creator also describes producing at 120 FPS before converting to 30 FPS; save that approach for later, preserving duration and audio synchronization during conversion.

The reference video at the top of this page is 960×720 (4:3) at 30 FPS. Choose the aspect ratio for your own video to suit its platform and purpose.