How to Create Custom AI Video?

Published · Updated

Before this: How to Create an Event?

HOW TOAI Assets -> AI Video -> Create

How to Create a Custom AI Video?

You design your own AI motion video with the Create New button in the Settings → AI Assets → Photo to AI Video tab: enter a name, AI model, video prompt, and optional generative visual prompt; It tests with an input photo if you want; You set the cover media (image or video) and the ring aperture option and save it. AI Assets require a Pro AI (or higher) subscription; Each test and final production consumes AI credits and runs on the server side.


i
Before you start
  1. Pro AI or Ultimate AI subscription, If AI Assets are locked, the Pro AI plan only lock card appears and is unlocked with Upgrade to Pro AI (before 3.6.0: 'Pro AI or higher required' / 'Upgrade to Pro AI')
  2. An event must be selected/open (the created video is added to the current event's AI Video library)
  3. Enough AI credits, each press of Test Effect and final production costs credits
  4. Internet connection, runs on test and production server side

Step by step

Step 1 · Open the Photo to AI Video tab

Settings → AI Assets → Photo to AI Video

Open the 'AI Assets' page from Settings. Click the Photo to AI Video tab from the tabs at the top (it sits alongside the AI Effects, AI Face Swaps and Video to AI Video tabs; before 3.4.2 this tab was called AI Video). If the page is locked with your subscription, you'll see the Pro AI plan only lock card; you can upgrade with the Upgrade to Pro AI button.

🔎 Items on the screen: Tab bar: AI Effects / AI Face Swaps / Photo to AI Video / Video to AI Video; Upgrade to Pro AI button when locked.

AI Assets page, Photo to AI Video tab selected: tab description below it, Surprise Effect key at top, AI Video Effects…
AI Assets page, Photo to AI Video tab selected: tab description below it, Surprise Effect key at top, AI Video Effects Library title, and Remove All + Browse Library + Create New buttons at top right.

Step 2 · Open empty modal with Create New (edit/clone difference)

AI Assets → Photo to AI Video

For a new video from scratch, click the Create New button in the top right, an empty rendering modal will open. (If you want to work with an existing card instead: clicking on a video card YOU created, anywhere except the round play button in the middle, opens the Edit & Customize modal and updating the effect in place; clicking on a video card that is NOT YOURS or preset/default opens the Clone & Customize modal, appends '(Copy)' to the name, and saves it as a NEW effect, the original remains unchanged.)

🔎 Items on the screen: Create New button; pencil (Edit) or duplicate (Duplicate) icons on the card; modal header Create a Video Effect / Edit & Customize / Clone & Customize respectively.

Opened Create a Video Effect modal: title + information badges, two preview boxes at the top (input photo and effect…
Opened Create a Video Effect modal: title + information badges, two preview boxes at the top (input photo and effect output), form fields at the bottom.

Step 3 · Read knowledge badges (credit + time)

Create AI Video, modal title

Examine the three information badges next to the modal title: the AI Motion Video label, the selected model's final production credits ('...Credits'), and the estimated production time (seconds, e.g. '~...s'). The credit badge is updated according to the AI Modelyou choose below.

🔎 On-screen items: AI Motion Video badge; credit badge (value + Credits); Time badge with clock icon.

Step 4 · Set input photo (default or your own image)

Create AI Video, left preview box

The left preview box is the INPUT photo that will be used during testing. By default, a sample photo is automatically uploaded. If you want to use your own photo, click the Upload Your Own Photo button at the bottom of the box and select an image from the native file selector that opens. While the photo you picked is being prepared and uploaded, the box shows Loading and the button is disabled. If there is no photo, it says No photo in the box.

🔎 On-screen elements: Left preview box; Upload Your Own Photo button; placeholder No photo when empty; Effect output placeholder in the idle case in the right box.

Two preview boxes side by side: input photo + Upload Your Own Photo button on the left, Effect output placeholder and…
Two preview boxes side by side: input photo + Upload Your Own Photo button on the left, Effect output placeholder and 'Test Effect (... Credits)' button on the right.

Step 5 · Enter effect name (mandatory)

Create AI Video, form

Type the name of the video in the Effect Name field. The field is required (with a red * next to it). Placeholder example: e.g. Neon Cyberpunk. If you leave it blank, This field is required warning will appear in the record.

🔎 Elements on the screen: Effect Name text box; e.g. Neon Cyberpunk placeholder; This field is required message in case of error.

Step 6 · Select AI Model (quality/cost is determined here)

Create AI Video, form

Select a model from the AI Model drop-down. Each option shows the model's name and credit cost in the form 'Model, N Credits'. When you change the model, the generation credit and Test Effect credit indicators at the top are updated as well.

3.7.0 and later: additional settings appear below the model, depending on the model you selected. Credits are calculated per second; the price of the length you chose is shown below the setting ('Longer videos cost more credits. This length costs:').

ModelAdditional settingsCredits per event shot
Hailuo (default)Resolution: 768p (6 or 10 s) / 1080p (6 s only) · Video Duration768p: 6 s 36, 10 s 60 · 1080p: 6 s 60
KlingVideo Duration: 5 or 10 s · Generate audio ('Adds sound to the video and doubles the credit cost.')5 s 65, 10 s 130; double with audio on
Activation KingVideo Duration: slider from 1 to 15 s9 per second
GenjutsuReference video + Resolution: 480p / 720pper second: 480p 55, 720p 120

Test Effect spends roughly half of this on every model.

Genjutsu (motion transfer): upload a 4-10 second clip with Choose a reference video. Description: 'The motion comes from this clip and the guest's photo provides the identity. The video must be 4-10 seconds; its length sets the price.' The price is calculated from the video's length and appears in seconds in the Billed row. If the clip is outside this range, it is not added: 'This video is {0} s long. Only 4-10 second videos can be used, so it was not added.' You can take the video out with Remove.

🔎 Items on the screen: AI Model drop-down list; option format 'Model, N Credits'; Activation King label (model name displayed in the interface).

AI Model drop-down menu is open: model options and credit cost are listed next to each.
AI Model drop-down menu is open: model options and credit cost are listed next to each.

Step 7 · Enter prompt (description of video: required)

Create AI Video, form

In the Prompt field, write the text that describes what the video/motion you want to create will look like. The field is required (*). Placeholder: Describe the effect you want to create.... This is the main directive of the video; If it is empty, no recording will be made. Below the Prompt box, 7 chips are listed under the Available personalization variables: label: {Name} {Number} {Color} {Interest} {Scene} {Weather} {CustomPrompt}. Clicking a chip pastes that variable at the cursor position; at capture time, the server replaces these placeholders with the real session data (the guest's name, their chosen color, interest, etc.). When you add a variable to the prompt, the Test values panel appears below the chips (Only used for Test, at the booth your guests enter these values.): only the fields used in the prompt (Name, Number, Color, Scene, Weather, Interests, Custom Prompt) are shown, pre-filled with sample values, and Test Effect runs with these values; you can also test without changing them. (3.4.0 and later)

🔎 Elements on the screen: Prompt multi-line text box (auto-growing); Describe the effect you want to create... placeholder; 7 personalization chips below it ({Name}…{CustomPrompt}); if the prompt contains a variable, the Test values panel.

The Prompt box (containing a variable such as {Name}) with the 7 personalization chips below it ({Name} {Number}…
The Prompt box (containing a variable such as {Name}) with the 7 personalization chips below it ({Name} {Number} {Color} {Interest} {Scene} {Weather} {CustomPrompt}), the Test values panel filled with sample values, and the Source Image menu below it

Step 8 · Source Image and Generative Image Prompt (optional pre-conversion)

Create AI Video, form

The Source Image menu below the Prompt field determines whether the photo is pre-processed before the video is generated. The default is None: No pre-processing, the captured photo is animated directly. If you choose AI Effect, you pick an effect from the AI effect library with the Select a AI Effect button (after selecting, the effect's name and Change appear); description: The selected AI Effect runs on the captured photo; its result becomes the video's source image. If you choose Prompt, the 'Generative Image Prompt' field opens; fill in this field, it says '(Optional)' next to the label and can be left blank. Placeholder: 'Describe how to transform the photo before the video is generated...'. The description of the field reads: 'If set, the photo is first transformed with AI using this prompt and the result is used as the video input.' So this field is for modifying the photo BEFORE video production; the actual video description is written in the Prompt field above. This field also has personalization chips; if you use a variable, the same Test values panel appears below it (the values are shared between the two fields).

🔎 On-screen elements: Source Image menu (None / Prompt / AI Effect); with Prompt selected, the 'Generative Image Prompt' textarea + '(Optional)' tag and hint text; with AI Effect selected, Select a AI Effect / Change.

Step 9 · Run Test Effect and watch the result (optional)

Create AI Video, right preview box

If you want to see the result before saving (not required), click the Test Effect button; The button writes the test cost in credits ('Test Effect (... Credits)'). The text Generating... and the elapsed seconds counter appear in the right box; when the job is finished, the result video autoplays in the box and also opens automatically in a full-screen preview window (muted, playing in a loop), here you see landscape videos, which are cropped in the small box, in full; Download saves it and Close takes you back to the form. The eye icon (Preview) at the top right of the result box reopens the full-screen preview; the download (Download) icon next to it saves the result to your device. (Full-screen preview: 3.5.1 and later.) If the test fails, the error is shown in the box.

🔎 On-screen items: 'Test Effect (... Credits)' button; Generating... + seconds counter; result video player + eye (Preview) and download (Download) icons; in the full-screen preview Download / Close; idle Effect output / error status placeholders.

While the generated test video is playing in the right preview box, the eye ( Preview ) and download icons at its top…
While the generated test video is playing in the right preview box, the eye (Preview) and download icons at its top right; 'Test Effect (... Credits)' at the bottom and Use as Cover buttons if necessary.

Step 10 · Set cover media (image OR video)

Create AI Video, Cover Media

Determine the cover that will appear on the card from the Cover Media section. Select an IMAGE or VIDEO from the local file chooser with the Pick cover image or video button (difference from AI Effects: here the video can also be a cover). Alternatively, you can cover the test video with the Use as Cover button that appears below the right box after a successful test. When the cover is set, the label becomes Cover image set; It is rolled back with Remove cover (Loading content... is shown when loading). Recommended size is Recommended 640x400px.. If you don't choose any cover and your test is successful, the last test video will be automatically made a cover.

🔎 Items on the screen: Pick cover image or video button; Remove cover button; Use as Cover button; Optional cover image or video / Cover image set tags; Recommended 640x400px. hint; Loading content....

Cover Media area: cover preview box + Pick cover image or video and (if cover) Remove cover buttons.
Cover Media area: cover preview box + Pick cover image or video and (if cover) Remove cover buttons.

Step 11 · Make Public (optional)

Create AI Video, form

If you turn on the Make It Public switch (the status becomes Enabled), this effect will become discoverable and available to others in Activation King's public gallery, as stated in the description. When it's off, the status says Disabled and the effect is only for you.

🔎 Items on the screen: Make It Public on/off switch; status text Enabled / Disabled; description below.

Step 12 · Save

Create AI Video, bottom bar

Hit the save button at the bottom: on new creation it says Save Effect, on edit it says Save Changes, on clone it says Save as New Effect. If Effect Name or Prompt is empty, no recording is made, missing fields are marked with This field is required and the screen scrolls to the first error. On successful save, the modal closes and the video is added to its activity's AI Video library. If you exit with Cancel, no changes will be saved.

🔎 Items on the screen: Cancel button; save button (Save Effect / Save Changes / Save as New Effect).

i
Tips

Credit indicators change depending on the selected AI Model: the top badge shows the final production cost, the value on the Test Effect button shows the testing cost.

Prompt differs from 'Generative Image Prompt': 'Prompt' identifies the video itself (mandatory); the 'Generative Image Prompt' is for converting the photo BEFORE the video (optional; it only appears when Prompt is selected in the Source Image menu).

If you do not select a cover, the result video of a successful test will be an automatic cover. If you don't want it, choose your own cover or say Remove cover.

The cover can be both visual and video; Pick cover image or video accepts both.

The Activation King model appears with this name to hide branding in the interface (it is a different model in the background).

The test requires the input photo to be uploaded to the server; You can test a newly uploaded photo as soon as it finishes uploading (there may be a short wait).

Selecting a cover file does not prevent recording; Uploading the cover to the cloud can be completed in the background after recording.

File selections (input photo and cover) are made with the device's native file selector.

!
Common mistakes

Attempting to save with Effect Name or Prompt left blank: This field is required exits and save is blocked.

Choosing a long duration and assuming the cost is the number on the model's label: credits are calculated per second; the actual price is shown in the line below Video Duration and in the credit badge at the top.

Turning on Generate audio on Kling when you don't need it: the credits double.

Thinking 'Productive Image Prompt' is mandatory or writing the actual video description here instead of 'Prompt': this field is optional and only pre-converts the photo.

Thinking the test is free: each press of Test Effect consumes as many credits as written on the button.

If the test fails, the box shows Test failed; if the request was caught by the content filter, the explanation 'This request was blocked by the content filter. Change the prompt or the input and try again.' appears below it. (In 3.8.0 and earlier the box showed 'TestFailed' in English.)

Thinking that you are editing a ready-made video that does not belong to you: in that case, the opened modal is Clone & Customizedir, '(Copy)' is added to its name and saved as a NEW effect; remains as original.

Overlooking that when Make It Public is on (Enabled) the effect will be visible to everyone.

Related topics

Adding Stock AI Video Creating AI Effect

Was this article helpful?

Not sure where to start? We can help.

Tell us about your project. Our team will help with the next step.

By submitting, you agree to our Terms and Privacy Policy.