Replace and correct assets

After source analysis, Settings maps original characters and locations to new ones. Correct identities before generating replacement assets in bulk.

Inspect mappings

Check each original character's name, identity and episode appearances. Cards without variants can be edited directly. Stacked cards contain costume variants; expand them and add outfits as needed.

New-asset prompts adapt the source to the target language context. Write a brief description and use Intelligent Optimization to develop it, or select library assets and uploaded references.

Check the mapping between original and replacement characters.

Check the mapping between original and replacement characters.

CoolShowAI

Expand variants and edit new character or location descriptions.

Expand variants and edit new character or location descriptions.

CoolShowAI

Merge duplicates

When one character is detected as several main appearances, prefer Merge duplicates (合并去重) to retain source associations and episode information. If the difference is clothing, organize it as a variant of the same main character.

Do not simply delete duplicates and continue. Deletion can remove source relationships and make later prompts misidentify characters.

Use asset correction to organize duplicate appearances and variants.

Use asset correction to organize duplicate appearances and variants.

CoolShowAI

Restore missing or deleted assets

1. Create the missing character or location.
2. Add it to the correct episode appearances.
3. Return to affected shots and insert the correct asset using @.
4. Check original character names in dialogue to avoid speaker mismatches.

Restoring an asset also requires episode association and shot references.

Restoring an asset also requires episode association and shot references.

Configure target-language voices

Intelligent voice generation uses character biographies to propose voice descriptions. Edit and audition them. English audition text helps evaluate an English-language production.

Prefer configuring voices before storyboard analysis. If added later, use intelligent linking and inspect actual shot references. A generated voice is not proof that every shot is correctly linked.

Generate and audition character voices in the target language.

Generate and audition character voices in the target language.

CoolShowAI

Before storyboard analysis

Every original character has the correct replacement; names are accurate; duplicates are merged; variants belong to the right character; missing assets are linked to relevant episodes; locations and props are complete; speaker voices are approved. Then begin shot analysis to reduce rework.

Deduplicate versus merge as an outfit

1. Open the card you want to retain and choose Deduplicate (合并去重) or Merge as outfit (合入变装). It becomes the target identity.
2. Select the other cards to process; verify the original person and episode associations.
3. Deduplicate: the selected duplicate cards are removed and the target inherits their episode information.
4. Merge as outfit: the selected images become outfits under the target, useful when one person was mistaken for several identities.
5. Expand the resulting outfits and check every episode association. Asset correction offers undo; if the merge was wrong, use the available undo and inspect the restored state.

The upper-right add controls can restore missing characters, scenes or props. Associate them with their episodes and reinsert their mentions in affected shots. Scene cards provide original/new comparison and original-scene output; asset images can be batch-downloaded.

Using English audition text for intelligent voices can improve subsequent English voice stability. Still check actual pronunciation, lip sync and speaker ownership in the generated clip.