Vitent Product Guide

Prepare assets and storyboards

Turn a saved screenplay into reusable visual references and structured clips. Review names, prompts, reference images, timing, actions, and linked assets before Production.

Updated September 10, 2026

Extract and generate reusable assets

Asset map production includes a set of core production processes, as well as several extended operations after asset completion. The core process needs to be completed in sequence; extended operations can be selected and used according to project needs.

Core production process

Step 1: Extract assets

After clicking "Confirm and Execute", the Agent will extract reusable assets based on the final script and write them into the current project asset library. After entering the "Assets" page, you can see that the assets of this episode are displayed in categories such as "Characters", "Scenes", "Props" and "Others". The specific asset names are also listed in the directory on the left.

English asset library with generated references for Milo, Nim, Heavenly Corner Store, and Silver Lantern
The asset page groups the four extracted assets and displays their generated reference images.
  • Start by checking that the assets cover key characters, scenes, and props.
  • Confirm again whether the asset name, description and classification are convenient for subsequent production and reuse.
  • If a key asset is obviously missing, you can submit modification comments in the Agent before continuing to generate the asset image.

Step 2: Generate asset prompts

After the asset structure is confirmed, let the Agent generate corresponding picture prompts for the characters, scenes and props. At this time, you can check whether the prompt accurately reflects the asset information, and modify the specific content through the Agent. After confirming the prompt, enter the asset image generation.

After generating the prompt, the system will display the specific content of the picture prompt, including visual style, asset information, appearance characteristics, layout and quality requirements, etc. Users can check whether the prompts are consistent with the expected description item by item; if they find that information is missing or the expression is inaccurate, they can modify the prompts first and then continue to generate pictures.

English Generate Image dialog with Milo's reviewed prompt, configuration, and credit estimate
Review the English prompt, reference images, model, size, ratio, quantity, and current credit estimate before submitting.

Step 3: Generate asset images

Open an asset variant and choose Generate Image. Confirm the reviewed prompt and generation settings, check the live credit estimate, and submit the task. Repeat this for the assets required by the episode. Generation runs in the background and the task list reports completion.

In this English-interface test, one reference image was generated for Milo, Nim, the Heavenly Corner Store, and the Silver Lantern. Each initial 1K, 1:1, Standard generation displayed an estimate of 12 credits and deducted 12 credits. A later regeneration dialog can show a different estimate when the existing image is included as a reference, so always use the amount shown at submission time.

After an image is completed, the asset card displays the selected result, which can be referenced in later storyboard and production work.

Four completed English asset cards for Heavenly Corner Store
The completed cards show Milo, Nim, the location reference sheet, and the empty Silver Lantern.

Operations after asset generation

View history and regenerate images

If you are not satisfied with the results of the current form, you can click "⋯" in the lower right corner of the asset image and select "Generate Picture" to produce a picture of the form again. Regeneration will not overwrite the original records, and new results will continue to be added to the generation history of the form.

Click "View History" in the asset form menu to view all pictures produced under that form. You can select a picture from the history record as the currently used picture, or you can reference, copy, move, download or delete historical pictures.

Click on any asset card to enter the asset details page. On the asset details page, you can view the current asset image, return to the assets in this set, edit asset information, or enter canvas editing to continue adding and adjusting materials.

Milo's generated reference sheet and English prompt in the asset preview
The image preview preserves the generated result, the exact English prompt, and generation metadata.

Managed asset form

After entering an asset, you can also add multiple forms to the same asset. For example, a character asset can contain adult, old or juvenile forms at the same time, and the corresponding form can be selected and used later according to the needs of the plot. When adding a new form, you can upload existing images or directly use AI to generate images.

Assets can set a default shape. The default form can usually be understood as the most frequently used and most important form of the character or asset in this episode; other forms are selected as needed in specific plots or scenes.

Set a character voice

Open the “⋯” menu on a character asset and choose Voice Settings to select or create a voice. Filter the voice library by attributes such as gender, age, and language. The selected voice appears beside the character name and follows that character into production.

To remove a selected voice, hover over the voice label beside the character name and click the “×”.

Edit in asset canvas

Click "Go to Canvas Edit" on the asset details page to enter the asset canvas. Use the import control to bring the asset's existing images and forms into the canvas for centralized review and editing.

The drop-down menu in the upper left corner of the asset canvas can switch canvases for different assets, such as switching between characters, scenes, and props. After switching, you can view the imported images and forms of the corresponding assets.

Build storyboards and enter Production

After the asset image is generated and confirmed, you can switch to "Storyboard" to generate the storyboard for this episode from the script, or you can let the Agent generate the storyboard based on the uploaded script. The storyboards will be organized by scenes and clips, and each clip will be broken down into shots and display information such as shot number, duration, scene, camera position/movement, subject, and action.

Audited English Storyboard with four clips and linked project assets
The regenerated Storyboard contains one scene and four clips of 13, 12, 14, and 8 seconds, with English shot descriptions and linked project assets.

After generating the storyboard, use “@” in the Agent input to reference a specific clip and request changes. You can split it into more shots or revise its action, pacing, and visual description.

If you only need to modify a certain paragraph of copy, you can also use the mouse to select the corresponding text, click the pop-up "Quote" button, bring the selected content into the Agent dialog box, and then make local adjustments. After the modification is completed, the Agent will generate a modification plan and update the corresponding storyboard after confirmation of execution.

The workbench Agent shown on the right assists with the current Drama project. Referencing a clip or selected text keeps the requested change scoped to the intended storyboard content.

  • Check that the storyboard accurately conveys scene sequence, camera action, and plot emphasis.
  • If the storyboard structure obviously deviates from the script, you can submit modification comments in Agent before continuing with production.
  • After confirming the storyboard, go to the "Production" page to process specific clips.

Open the Production page to enter the infinite canvas and select the current clip in the Timeline. Clip numbers use the format “scene.clip”: 1.1 is the first clip in Scene 1, while 2.2 is the second clip in Scene 2.

Select Import All. In the verified English-interface run, Vitent imported three reference nodes — Heavenly Corner Store, Milo, and Silver Lantern — created a video node, connected the references, and added the current clip's three-shot storyboard to the prompt.

English Production canvas after Import All connects three assets to a video node
Import All brings the clip's approved assets and storyboard prompt into the production canvas.

Check the imported references and prompt

Before generating, verify that every connected reference belongs to the current clip and that the shot text matches the intended story. Add or remove references when the prompt needs a different character, location, prop, image, video, or audio source.

Add canvas nodes manually

Use the plus button on the left side of the canvas when the clip needs an additional text, image, video, or audio node. Connect only the sources that should influence the selected generation node.