Chapter 27 covers two specialized AIUNIFY PROS SUITE AI Marketing workspaces:
Both create AI-generated Video content and both support reusable Product and Avatar references, but their production workflows are different.
Video Generator is presented as a Product Video Generator designed to turn a Product, Avatar, or described scene into a single cinematic marketing Video.
Reels Generator is a multi-stage short-form production workflow that can generate a script, voiceover, music information, individual rendered scenes, and a final Reel.
The major distinction is:
Video Generator → One generated marketing Video task
Reels Generator → Multi-stage short-form production pipeline
Navigate to:
Sidebar → AI Marketing → Video Generator
The current route is:
/ai-marketing/video-generator
The Sidebar marks the Tool as New.
The actual workspace is titled:
Product Video Generator
with the description:
“Turn static product images into cinematic marketing videos.”
Video Generator checks whether the subscriber is authenticated.
If the session is unavailable, the current page redirects toward:
/login.
Video Generator does not use the Standard AI Generator interface.
It uses the specialized Marketing media service and provides:
The basic workflow is:
Select Product and/or Avatar
↓
Describe Movement
↓
Choose:
Mode
Resolution
Aspect Ratio
Duration
↓
Generate Video
↓
Wait for rendering
↓
Preview
↓
Download or Regenerate
↓
Review Video Gallery
The first reusable creative input is:
Select Product Reference
The same Product Picker described in Chapters 25 and 26 is used here.
If the Product does not already exist in your reference library, use:
Upload Product
The current Product Picker accepts Image files and stores them for reuse across supported Marketing workflows.
The current Video Generator's handleGenerate function validates only the Movement Prompt.
It does not explicitly block generation when no Product or Avatar has been selected.
However, because the workspace is designed as a Product Video generator, selecting an appropriate reference is recommended for predictable results.
The backend's acceptance of every no-reference request is outside the sanitized source.
Video Generator also provides:
Select AI Avatar
using the shared Avatar Picker.
When only a Product is selected, the generation Prompt is prefixed with:
“The product ([Product Name]) in the scene.”
The Product's stored Image URL becomes the primary reference Image sent to the Video service.
When only an Avatar is selected, the Prompt is prefixed with:
“An AI avatar ([Avatar Name]) in the scene.”
The Avatar Image is supplied as the current visual reference.
When both are selected, the Prompt is prefixed with:
“An AI avatar ([Avatar Name]) interacting with a product ([Product Name]).”
When both Product and Avatar exist:
imageUrl uses the Product Image.avatarUrl separately sends the Avatar Image.The rewritten Prompt also identifies both by Name.
For a spokesperson-style Product Video, the strongest current workflow is:
Product Reference + AI Avatar + Movement Prompt
For example:
“The spokesperson holds the Product toward the camera while the camera slowly circles from left to right. Use premium studio lighting, realistic reflections, and a slow cinematic push-in.”
The Prompt section is labeled:
Step 2: Describe Movement.
The current example is:
“E.g. The camera circles the product smoothly with a slow zoom in, showing cinematic highlights and reflections...”
If the field is empty, the current message is:
Please enter a movement prompt.
Image Generator focuses heavily on what a scene should look like.
Video Generator should additionally explain:
The interface itself suggests terms such as:
Examples include:
For example:
“Make a cool Video.”
provides very little motion guidance.
For example:
“Begin with a wide shot of the Product on a dark marble pedestal. Slowly orbit clockwise while pushing closer. Add soft light sweeps across the surface, then finish with a centered hero shot and a subtle slow-motion glow.”
Video Settings includes:
Mode.
The current choices are:
The current default is:
Normal.
Use Normal when you want the service's standard/balanced Video-generation behavior.
Use Fun when you want to request a more playful creative direction.
The exact provider-side transformation associated with fun is outside the user-operations source.
Use Spicy when you want a more dynamic variant.
Again, the current client sends the value but the provider's exact implementation is not part of the sanitized source.
Unlike Chapter 26's Image Style selector, Video Generator's selected Mode is included in the generation request.
Therefore, Mode is source-verified as a real request parameter.
Video Generator provides:
Resolution.
The current choices are:
The current default is:
480p.
The current Product Video Generator does not expose:
as subscriber-selectable resolutions.
Unlike the Image Generator's ratio-like Resolution label, Product Video sends the actual string:
480p720pas its Resolution setting.
The current Video Generator offers:
Aspect Ratio.
The current choices are:
The default is:
16:9.
Useful for:
Useful for:
Useful for square social placements and general feed creative.
These provide additional landscape/portrait composition options.
The selected value is submitted as:
aspect_ratio.
Video Generator includes a Duration slider.
The current range is:
6 to 30 seconds
in:
1-second increments.
The current default is:
6 seconds.
The label changes dynamically to:
Duration ([number]s).
The selected slider value is submitted to the Marketing service as a string representation of seconds.
The primary action is:
Generate Video.
While processing, it changes to:
Directing Video...
The current generation request includes:
action: generate-videoimageUrlProduct Video posts to:
/api/marketing
rather than the normal text-generation AI stream.
The current Product Video component does not retrieve or submit the shared PROS SUITE:
selectedModel.
Therefore:
Changing the normal Header language Model is not source-verified to affect Product Video generation.
The Product Video request also does not append the normal provider API-Key headers described in Chapter 11.
The specialized Marketing service controls this generation path from the subscriber-client perspective.
There is no current subscriber-facing Video Model selection field inside the workspace.
A successful Video request initially returns a:
taskId
rather than the completed Video itself.
The current notification says:
“Video generation started. This may take 2-3 minutes.”
Once a Task ID exists, the client checks status approximately every:
5 seconds.
When the service reports:
success
the client:
resultUrlThe exact current success message is:
Video generated successfully!
If the service reports failure, the client stops processing and displays the returned Error or:
Generation failed.
The client specifically checks for an Error containing:
credits insufficient.
In a demo-domain condition, that becomes:
In the demo, this feature is disabled.
This demo-specific message should not be interpreted as the standard production behavior.
If the service cannot start the Video task, the displayed message is the returned Error or a message based on:
Failed to start generation.
The fallback is:
Video generation request failed.
While processing, the Preview displays:
Rendering Cinema...
The current interface states:
“Kie.ai is processing your high-end marketing video. This usually takes 2 minutes.”
The rendering screen includes a Progress animation configured to fill across:
120 seconds.
The bar is not connected to a returned provider percentage.
Actual completion is determined by the 5-second polling workflow.
Therefore:
The 120-second bar is a visual time-based animation, not an exact completion measurement.
Use the success/failure state as the authoritative completion indicator.
Product Video currently does not provide:
once the generation task has started.
When handleGenerate begins, the previous current Video is removed from the primary result state.
Preserve important output before creating a replacement.
Before a Video exists, the right panel displays:
Video Preview
and:
“Select an asset above and describe the movement to see the magic happen.”
A successful Video is displayed in the right panel as an HTML Video element.
The main result Preview is configured with:
but its visible playback controls are disabled.
The main result's Video element uses:
controls={false}.
If you need full playback controls, open the Video through Asset Viewer/Video Gallery.
Unlike Gallery thumbnail Videos, the main result Preview does not include a muted setting in the current source.
Browser autoplay behavior can therefore vary depending on normal browser media policies.
The top-right result controls include a Download button.
The current result is requested with a Filename in this pattern:
ai-suite-video-[timestamp].mp4.
The direct result Download uses:
.mp4.
The current Download control displays:
Download started.
As documented in Chapter 25, the shared Download helper can open the remote media URL in a new tab if the normal proxy-based download fails.
The second current-result control uses a Refresh icon.
It calls the same full:
handleGenerate
workflow.
It does not edit the existing Video.
It submits another complete Video-generation request using the current inputs.
Before Regenerating:
The main result container attempts to open Asset Viewer when clicked.
Like Chapter 26's Image Generator, the current Video result click sets:
viewerIndex = 0
while Asset Viewer receives the recentVideos array.
Therefore, the current Preview relies on Gallery ordering rather than directly giving Asset Viewer the resultVideo URL.
Use:
Video Gallery → Locate the correct Video → Open that Card
rather than relying on the current result's index-zero Viewer shortcut.
Below the generator is:
Video Gallery.
When Video Generator opens, it requests:
list-assets&type=video.
This establishes a persistent type-specific Video Asset collection.
A successful generation automatically triggers another Video-history fetch.
When no Video assets are returned, the exact current message is:
No video history found.
Each saved Video appears using the shared MarketingAssetCard.
That provides:
Shared Video Asset Cards:
while used as Gallery thumbnails.
Select the Card or View action to open Asset Viewer.
Inside Asset Viewer, Video assets are opened with playback controls.
This is the better environment for reviewing the complete media.
MarketingAssetCard uses a pattern such as:
marketing-video-[first 8 characters of asset ID].mp4
for historical Video downloads.
The Gallery Card's Share-style button copies the current Video Asset URL.
The exact confirmation is:
Link copied to clipboard.
The Video Generator does not pass a Delete handler into its MarketingAssetCards.
Therefore, no Delete control is currently available from this Gallery.
There is no source-verified:
Rename Video
control.
Video Gallery currently has no Search field.
There are no subscriber-facing:
inside Video Gallery.
The current Gallery does not provide:
The generic PROS SUITE feature catalogue currently lists:
Marketing Video Gen — 100 Tokens.
The Product Video page itself does not display a fixed 100-Token price or perform an explicit client-side Token deduction.
Therefore:
100 Tokens is the current catalogue/default reference, not proof that every production Video generation is permanently billed exactly 100 Tokens.
Final server-side accounting is outside the sanitized user bundle.
Because the media path can report:
credits insufficient
provider availability/credits should not automatically be treated as the same accounting mechanism as the subscriber's PROS SUITE Monthly Tokens.
The generator does not display:
inside its creation Card.
The success path refreshes Video Gallery but does not call the shared account refreshUser() function in this client.
Do not assume the Sidebar balance visibly updates at the exact moment Video generation finishes.
Before generating, decide whether the Video is intended for:
Then choose the appropriate Aspect Ratio.
A 6-second Video can work well when you need:
Longer durations should be used when the scene genuinely requires additional time.
For example:
“Start with a wide shot, slowly orbit around the Product, then finish with a close centered hero shot.”
This gives the generator a clearer temporal structure.
For commercial Product work, describe:
to reduce unwanted visual distortion.
Check:
throughout the Video rather than only at the first frame.
When using a spokesperson Avatar, check whether:
between frames.
Do not rely only on Gallery URLs for production storage.
Maintain an external copy of approved Marketing Videos.
If nothing starts, enter a Movement Prompt.
The exact validation is:
Please enter a movement prompt.
The client does not block generation solely because no Product/Avatar is selected.
However, if the Video service rejects the request or produces an unsuitable result, choose the intended reference asset and retry.
Verify Aspect Ratio before generating.
Regeneration is required to create another Video with a different requested ratio.
The current Resolution choices stop at:
720p (HD).
The 120-second bar is not actual task progress.
Wait for:
Video generated successfully!
or the failure state.
No current Stop/Cancel action is source-verified.
Navigate to:
Sidebar → AI Marketing → Reels Generator
The route is:
/ai-marketing/reel-generator
and the Sidebar marks it as New.
The page is titled:
AI Reel Generator
with:
“Generate viral short-form videos with AI.”
The page verifies authentication and redirects an unauthenticated subscriber toward Login.
Unlike Product Video's single Video task, Reels supports a production record containing information such as:
The current Progress calculation explicitly checks those fields.
The high-level workflow is:
Your Vision
↓
Optional:
Product Reference
AI Avatar
↓
Settings:
Mode
Aspect Ratio
Resolution
Duration
↓
Generate Viral Reel
↓
Script/production pipeline
↓
Scene Rendering
↓
Final Reel
↓
Preview Final
↓
Reel Archive
The main Prompt area is labeled:
Step 1: Your Vision.
The current example is:
“Describe your reel vision... e.g. 'A high-energy promotional reel for a new luxury skincare line.'”
If the Prompt is blank, the exact current notification is:
Please enter a reel topic or product description.
Unlike Product Video's movement-focused Prompt, the Reel Prompt can describe the broader production concept.
Useful information includes:
“Create a high-energy 15-second vertical Reel introducing AIUNIFY to sales organizations. Open with a fast business-problem hook, transition into three scenes showing CRM, AI automation, and team collaboration, and end with a premium Built for Business call to action.”
Reels Generator includes the shared:
Select Product Reference
workflow.
It also includes:
Select AI Avatar.
When an Avatar is selected, Reels appends:
“Include the person shown in the provided reference photo as the main character.”
When a Product is selected, Reels appends:
“Feature the product shown in the provided reference photo prominently.”
When both are present, it additionally says:
“The person should be naturally interacting with or showcasing the product.”
Reels builds a:
filesUrl
array.
It can contain:
and sends the entire array in the Reel-generation request.
Product Video uses dedicated:
imageUrlavatarUrlfields.
Reels sends the selected references together in:
options.filesUrl.
The upper-right Header includes:
Settings.
The current showSettings state begins as:
true.
The advanced settings are therefore initially displayed.
Select:
Settings
to hide or reopen the settings area.
Collapsing it does not reset the selected values.
The first advanced control is:
Mode.
Current options are:
The default is:
Normal.
The selected Mode is included in:
options.mode.
The next control is:
Aspect Ratio.
The current choices are:
The current default is:
9:16.
This fits the Tool's short-form/vertical orientation.
Even though 9:16 is the default, the current interface allows all five ratios above.
Reels provides:
Resolution.
The current options are:
Default:
480p.
The current UI does not expose:
for Reels.
The current Duration control is labeled:
Duration (6-30s).
The current allowed range is:
6 to 30 seconds.
Default:
6 seconds.
Unlike Product Video's slider, Reel Duration is entered through a numeric field.
The current input logic limits values to:
and falls back to 6 when necessary.
Reels Generator creates a random internal:
seed
when the page component initializes.
That seed is included in:
options.seed.
There is no visible:
Seed
field in the current Reel Settings.
The value exists internally but is not user-configurable through the interface.
The current state initializes once when the component mounts and is not changed by handleGenerate.
Therefore, repeated Reel generations during the same mounted session submit the same internally generated seed unless the page is recreated.
The exact provider effect of that seed is outside the client source.
The primary action is:
Generate Viral Reel.
During processing, it changes to:
Producing Reel...
Below the main button, the current interface states:
“Uses Grok for high-performance cinematic video rendering.”
This is the current user-facing provider/engine description for the Reel workflow.
The Reels component does not retrieve or submit the shared PROS SUITE:
selectedModel.
Therefore, changing the Header language Model is not source-verified to affect this Reel-generation pipeline.
The Reels request sends standard JSON headers and does not append the normal subscriber provider API-Key headers.
The specialized Marketing service controls this operation from the client perspective.
When a new generation begins:
reelData is cleared.
The new production replaces the current right-side status/output state.
The current Reel request sends:
action: generate-reelThe service returns a:
reelId
when generation begins successfully.
The exact informational message is:
“AI Reel generation started. This involves script, voiceover, and scene creation.”
The client expects the production data may contain:
voiceover_url.
The client also expects:
music_url.
Despite the production pipeline supporting a voiceover result, the current Reels interface does not expose:
The current UI does not provide a pre-generation field for manually writing or editing the generated voiceover script.
There is no user-facing setting for:
in the current Reels source.
Although the imported UI icons include text-related components, the rendered subscriber interface does not expose:
as current controls.
The current Reel Settings do not provide a field for:
Number of Scenes.
Scene construction is handled by the production pipeline.
While a Reel ID exists and the generation state is active, the client polls:
poll-reel
approximately every:
5 seconds.
Every polling response is stored into:
reelData.
This allows the interface to display incremental production progress before final completion.
When status becomes:
success
the client:
The exact success notification is:
AI Reel generated successfully!
If the status becomes:
failed
the exact current notification is:
Generation failed.
Unlike Product Video, the current failed Reel polling branch does not display data.error.
It uses the generic:
Generation failed.
If the initial request fails before a Reel ID is created, the UI displays the returned Error or:
Failed to start generation ([status]).
The fallback is:
Reel generation request failed.
The right panel becomes:
Production Progress.
The status Header also displays:
Live Generation.
Unlike Product Video's simple time animation, Reels calculates its percentage from returned production artifacts.
If reelData exists, the progress calculation begins with:
10%
representing the pipeline having started.
When a script is available:
+30%
is added.
When voiceover_url exists:
+20%
is added.
When music_url exists:
+10%
is added.
The remaining scene portion is based on:
Completed scenes ÷ Total scenes
where a completed scene is one containing a:
video_url.
If the Reel is not yet in success status, progress is capped at:
99%.
Only successful completion is forced to:
100%.
The percentage reflects returned production components rather than merely elapsed animation time.
However, it is still a client-calculated progress model—not a provider-supplied universal rendering percentage.
The visible status step is:
Scene Rendering.
While incomplete, the current technical status label shown beneath it is:
RENDERING_VIDEO...
Once every scene in the script has a video_url, the visible status becomes:
COMPLETED.
Although voiceover and music contribute to the percentage, the current rendered status-step list contains only:
Scene Rendering.
There are no separate visible Cards titled:
When a script exists, the right panel adds:
Production Scenes.
Each scene is numbered:
1, 2, 3, etc.
Each Scene Card displays:
scene.text
inside quotation marks.
There is no user-facing control to:
during the current Reel production process.
When a Scene has its own video_url, a small Video thumbnail appears.
Those Scene Video thumbnails are:
The current Scene list does not provide a subscriber control to Download individual Scene Videos.
There is no current:
Regenerate Scene
action.
When the production status is successful, the final panel displays:
Reel is Ready!
with:
“Final export complete.”
The success panel provides:
Preview Final.
Preview Final sets:
viewerIndex = 0
and opens Asset Viewer.
However, Asset Viewer receives the current:
recentReels
Archive mapped into Video assets rather than the current reelData object directly.
The current implementation assumes the newly completed Reel is available at the first position in the refreshed Archive.
If Archive ordering or refresh timing does not place it there, Preview Final could open a different Reel.
Use:
Reel Archive
↓
Locate the correct Reel by:
↓
Select:
PLAY
Before generation, the right panel displays:
Production Studio
with:
“Enter a vision on the left to start the multi-step AI production pipeline. Your masterpiece will appear here.”
Like Product Video, Reels does not currently provide:
once the production begins.
There is no dedicated:
Retry Reel
control after completion.
To create another version, select:
Generate Viral Reel
again using the current or revised inputs.
A new Generate action clears current reelData and starts another Reel ID.
It is a new production operation rather than an edit of the finished Reel.
Below the production workspace is:
Reel Archive
with:
“Access your previously generated viral reels.”
The Header displays:
[number] Items.
When the page loads, Reels Generator requests:
list-reels
and stores the returned Reel records.
When a Reel reaches success, the generator calls:
fetchRecentReels().
The final production can therefore appear in the Archive automatically.
Every Reel Archive Card uses a fixed visual aspect container of:
9:16.
Because Reels Generator allows:
but Archive Cards are always displayed in a 9:16 container using object-cover, non-vertical Reel previews can appear cropped in the Archive.
If an Archive thumbnail appears cropped, select:
PLAY
and review the Reel in Asset Viewer.
Do not assume the Video file itself is cropped merely because the vertical Archive Card is.
When the pointer enters an Archive Card, the Reel Video is played.
When the pointer leaves:
Archive thumbnails use:
muted
and loop inline.
If a Reel does not yet have result_url, the Card displays a placeholder.
For a Reel whose status is processing, the placeholder displays:
Generating...
For another Reel without a final URL, the fallback is:
No Preview.
The hover overlay displays the Reel's stored Prompt.
If stored Reel metadata contains:
metadata.options.mode
the Archive Card displays the Mode in an uppercase badge.
Select:
PLAY
to open the chosen Reel in Asset Viewer.
The Archive Card includes a Download action.
The current pattern is:
ai-reel-[reel ID].mp4.
After the action is requested, the current notification is:
Download started.
The final downloadable Reel uses:
.mp4.
The success banner itself provides:
Preview Final
but does not render a Download control.
The current verified Download action is located on the Reel Archive Card.
After successful completion:
Before passing Reels into Asset Viewer, the page maps:
result_url → urltype → videoThis allows the shared Marketing Asset Viewer to treat Reels as Video assets.
If no saved Reels are available, the interface displays:
Archive is Empty
and:
“Your generated reels will be saved here for easy sharing and downloads.”
The custom Reel Archive Cards do not provide a Delete action.
There is no source-verified:
Rename Reel
control.
No Search field exists in Reel Archive.
There is no user-facing Sort selector.
There is no user-facing filtering by:
No current:
is established.
Unlike MarketingAssetCard, the custom Reel Archive Card does not include a Copy Link/Share URL button.
Its current actions are:
The Reel's generated scene/script information cannot currently be downloaded separately through this page.
The source recognizes voiceover_url for production progress, but no subscriber-facing control downloads the voiceover separately.
Likewise, music_url can exist in Reel production state but there is no current standalone Music Download action.
The current Reels interface does not expose:
The current page does not source-verify:
Reels Generator does not expose a Video-editing timeline where subscribers can:
There is no current:
Replace Scene
control.
Reels Generator does not provide a user-facing workflow for uploading an existing MP4 and asking AI to remix/edit it.
There is no current voice-cloning or voice-upload field in this Reel workspace.
There is no source-verified user control for uploading your own Music track.
The current Reel page does not display a fixed Token price beside:
Generate Viral Reel.
The retrieved AI Marketing Token catalogue explicitly lists Marketing Video Gen — 100 Tokens along with Image, Avatar, Logo, and other Tools, but the current Reels page is a separate reel-generator workflow and does not display a Reel-specific Token cost.
Therefore:
Do not automatically apply the Marketing Video 100-Token catalogue value to Reels Generator.
A fixed Reel Token charge is not source-verified in this User Operations bundle.
This distinction is reinforced by its use of:
action: generate-reel
rather than:
action: generate-video.
The current interface does not provide separate prices for:
This manual does not invent them.
Since Vertical 9:16 is the current default, it is the natural choice for conventional mobile short-form content.
Use another ratio only when your destination requires it.
For a short Reel, use the Prompt to specify what happens immediately.
Example:
“Open with the problem: sales teams using too many disconnected systems.”
For example:
Scene 1 — Problem
Scene 2 — Solution
Scene 3 — Product Benefit
Scene 4 — Call to Action
Even though you cannot manually control Scene count, this structure gives the AI clearer production direction.
Examples:
If written CTA accuracy matters, inspect the final Video carefully.
There is no voice selector, but you can still describe the intended production tone in the broader Reel Prompt.
For example:
“Use confident, professional, energetic narration appropriate for enterprise sales leaders.”
Whether every provider detail follows that request is generative and should be verified.
No Music selector exists.
If Music style matters, include direction such as:
“Use modern, confident cinematic technology music with an energetic but professional tone.”
Then verify the final output.
Use Production Scenes to inspect the current scene wording as it becomes available.
Look for:
After completion:
Archive → Download → Rename → Store with the campaign
rather than relying only on the in-app Archive.
The Archive thumbnail is muted.
Open the complete Video through Asset Viewer before approving:
Because Archive Cards are always 9:16, use Asset Viewer to judge actual composition for:
outputs.
If nothing starts, enter a Reel topic or Product Description.
The exact message is:
Please enter a reel topic or product description.
No voice selector is currently rendered.
Voiceover is part of the production pipeline, not a subscriber-configurable panel in this version.
No Music selector is available in the current source.
Production Scene text is visible but not editable.
A new generation is required if you want the AI to build another version.
No Scene timeline/reorder control is currently established.
The current Duration input is limited to:
30 seconds.
The maximum current visible Resolution is:
720p.
The client intentionally caps incomplete production at 99%.
It changes to 100% only when status becomes:
success.
Production Scenes can appear gradually as Scene Video URLs become available.
Continue waiting for the final status.
A Reel without a result_url displays:
No Preview
or:
Generating...
depending on its status.
The Archive uses a fixed 9:16 Card.
Open the Reel through PLAY/Asset Viewer to inspect its actual requested ratio.
Use the specific Reel Archive Card.
The current Preview Final action opens Viewer index zero of the Archive rather than directly supplying current reelData.
Use the Reel Archive's Download control.
The success banner provides Preview Final, not Download.
No Delete action is established in Reel Archive.
The custom Archive Card does not expose Copy Link.
Download the MP4 for reliable sharing.
There is no current Cancel/Stop action.
Before generating a Product Video:
After completion:
Before selecting Generate Viral Reel:
During production:
After success:
| CapabilityProduct VideoReels Generator | ||
| Route | /ai-marketing/video-generator | /ai-marketing/reel-generator |
| Main purpose | Single cinematic marketing Video | Multi-stage short-form Reel |
| Default Ratio | 16:9 | 9:16 |
| Product Reference | Yes | Yes |
| Avatar Reference | Yes | Yes |
| Product + Avatar | Yes | Yes |
| Prompt Type | Movement-focused | Full Reel vision/topic |
| Modes | Normal/Fun/Spicy | Normal/Fun/Spicy |
| Resolution | 480p/720p | 480p/720p |
| Ratios | 5 | 5 |
| Duration | 6–30 sec slider | 6–30 sec numeric field |
| Internal Seed | Not exposed in page | Yes, internally generated |
| Script Pipeline | Not established | Yes |
| Voiceover Pipeline | Not established | Source expects voiceover URL |
| Music Pipeline | Not established | Source expects music URL |
| Scene Progress | No | Yes |
| Final MP4 | Yes | Yes |
| Persistent History | Video Gallery | Reel Archive |
| Copy Asset URL | Yes via Gallery Card | No current Archive action |
| Delete | No | No |
| Cancel Generation | No | No |
| SettingProduct VideoReels | ||
| Mode | Normal | Normal |
| Resolution | 480p | 480p |
| Aspect Ratio | 16:9 | 9:16 |
| Duration | 6 sec | 6 sec |
| Seed | No visible setting | Random internal seed |
| Settings Panel | Always visible | Can be collapsed |
Both current Tools support:
Product Video labels 9:16 as Portrait, while Reels labels it Vertical.
Both current Tools expose:
Neither currently exposes 1080p or 4K through the subscriber UI.
The source verifies:
The current source does not establish:
The source verifies:
generate-reelThe current source does not establish subscriber controls for:
AIUNIFY PROS SUITE provides two distinct AI Video workflows inside the AI Marketing section.
The first is:
Video Generator
at:
/ai-marketing/video-generator
The actual workspace is titled:
Product Video Generator
with:
“Turn static product images into cinematic marketing videos.”
Product Video supports reusable Product and Avatar references. When a Product is selected, it can become the primary visual reference. When an Avatar is selected, its Image is also passed into the request, and the Prompt is automatically expanded to describe the selected subjects.
Subscribers then use:
Step 2: Describe Movement
and select:
All of these Video settings are actually submitted in the current generation request.
Generation runs asynchronously. After the initial request returns a Task ID, the client polls approximately every five seconds. During processing, the workspace says:
Rendering Cinema...
and:
“Kie.ai is processing your high-end marketing video. This usually takes 2 minutes.”
The visible 120-second Progress animation is time-based rather than linked to actual provider completion.
A successful Product Video can be downloaded as an MP4 using a filename such as:
ai-suite-video-[timestamp].mp4
and the Video Generator maintains a persistent:
Video Gallery.
The second workflow is:
Reels Generator
at:
/ai-marketing/reel-generator.
Its workspace is titled:
AI Reel Generator
with:
“Generate viral short-form videos with AI.”
Reels defaults to:
and supports the same five output ratios and the same 6–30-second duration range as Product Video.
Its production model is more complex. After a successful start, the interface states:
“AI Reel generation started. This involves script, voiceover, and scene creation.”
The client calculates production progress from the availability of:
and only displays 100% when final status becomes successful.
Subscribers can also watch Production Scenes appear with Scene text and small rendered Video previews. However, those Scenes are currently read-only: no Edit, Reorder, Delete, individual Download, or Scene Regenerate controls are source-verified.
When complete, Reels displays:
Reel is Ready!
Final export complete.
and provides:
Preview Final.
Final Reels are stored in:
Reel Archive
where subscribers can select:
PLAY
or Download an MP4 using:
ai-reel-[reel ID].mp4.
One important display limitation is that Reel Archive Cards always use a 9:16 shape even though the generator can produce several different ratios. Non-vertical Reels can therefore appear cropped in their Archive thumbnail, and should be reviewed through Asset Viewer.
Another important distinction is that the production data can include Voiceover and Music, but the current subscriber interface does not provide direct controls to select the Voice, upload Music, adjust Audio levels, edit Captions, or edit the generated production Script.
Finally, the generic PROS SUITE feature catalogue identifies Marketing Video Gen with a 100-Token reference value, but Reels Generator is a separate generate-reel workflow and does not display a Reel-specific fixed Token cost. Accordingly, this manual does not apply the Product Video 100-Token reference automatically to Reel generation.
The recommended operating model is:
For Product Video
Select Product/Avatar → Describe Movement → Choose Mode, Resolution, Ratio, and Duration → Generate → Wait for True Task Completion → Inspect the Full Video → Download → Use Video Gallery for History
For Reels
Define the Short-Form Vision → Select Product/Avatar References → Choose Mode, Ratio, Resolution, and Duration → Generate Viral Reel → Monitor Script and Scene Production → Wait for Reel Is Ready → Preview the Final Reel → Review Voice, Music, Scenes, and Branding → Download from Reel Archive → Preserve the approved MP4 externally