image-text-replace

Image Text Replace

Overview

Perform text content replacement on an existing image: simple precise text replacement (all four conditions satisfied: standard print font/single location/≤6 characters/character count unchanged) and complex text replacement (stylized text/handwriting/stylized/multiple locations/character count changes).

API Mapping

• Simple precise text replacement: simple text replacement path.
• Complex text replacement: prioritize praline_pro, and on failure degrade in order to praline_lite、mint_edit。

Dependencies

• meitu-cli: >=2.0.6
• credentials：CONFIG AKSK → meitu tools update; EXEC AKSK → actual execution (see root CONFIG.md）
• environment variables：MEITU_OPENAPI_TOOL_TASK_MODE=command

Path aliases:$VISUAL = {OPENCLAW_HOME}/workspace/visual/

Core Workflow

Preflight → Execute → Deliver

Preflight

1. meitu --version ≥ 2.0.6
2. Already ran with CONFIG AKSK meitu tools update
3. current AKSK = EXEC,MEITU_OPENAPI_TOOL_TASK_MODE=command
4. output_dir：openclaw.yaml → ./output/ ｜else → $VISUAL/output/image-text-replace/；mkdir -p

Execute

Trigger signals / routing rules

Core dimensions:four-condition check(must all be satisfied simultaneously):

1. Font type: standard print font (Songti/Heiti/Kaiti/Rounded/Serif/Sans-serif/Monospace)
2. Edit location: modify only one place of text
3. Character length: the number of characters in the modified text ≤ 6
4. Character count identity: original character count = modified character count

Scenario	Judgment	Route
All four conditions satisfied	Standard print font + single location + ≤6 characters + character count unchanged	Simple text replacement path → praline_pro → praline_lite → mint_edit
Stylized text / handwriting / stylized	Decorative lettering, handwriting, stylized	praline_pro → praline_lite → mint_edit
Multiple edits	Change multiple places	praline_pro → praline_lite → mint_edit
Character count >6 or character count changes	Long text / 3 characters→5 characters	praline_pro → praline_lite → mint_edit

Parameter definitions

Parameter	Type	Required	Description
image_url	STRING	Yes	Image address. If missing → prompt "Please provide the image that needs text replacement"
source_words	STRING	Conditionally required	Original text content. Required for simple text replacement; for complex text replacement it can be automatically extracted from prompt automatically extracted
target_words	STRING	Conditionally required	Target text content. Required for simple text replacement
prompt	STRING	No	Text replacement description; for complex text replacement, automatically generate an English text replacement instruction based on source_words / target_words automatically generate an English text replacement instruction

The simple text replacement path directly passes image_url、source_words and target_words; complex text replacement uses praline_pro、praline_lite or mint_edit, and constructs an English text replacement instruction based on source_words/target_words .

When the target text contains Chinese,prompt the target text inremains as-is and is not translated。

Tool invocation

meitu image-text-replace --image_url <url> --source_words "<原文>" --target_words "<新文>" --json --download-dir {output_dir} --skill_name skill_image-text-replace

Error degradation

Scenario	Handling method
image_url missing	Prompt "Please provide the image that needs text replacement"
image_url inaccessible	Directly return an invalid image link error, do not retry
source_words / target_words missing	Prompt "Please tell me which text to change to what"
source_words not found in the image	Return the original image and inform that no matching text was found
Character count change / styled text / multiple places	Do not use simple text replacement; directly use praline_pro
Simple text replacement path call failed	Automatically upgrade to praline_pro
praline_pro call failed	Automatically downgrade to praline_lite
praline_lite call failed	Automatically downgrade to mint_edit
mint_edit call failed	Return error
User wants to edit multiple images at once	Prompt to process one image at a time
User did not provide font information	Judge as standard print font, try simple API first; automatically upgrade on failure
Content compliance interception	Directly return compliance prompt, no retry, no downgrade

Deliver

• Use output_dir parsed by Preflight
• From downloaded_files[0].saved_path read the downloaded file path
• mv {downloaded_files[0].saved_path} {output_dir}/{YYYY-MM-DD}_{descriptive}_image-text-replace.{ext}

Output

• Format：image
• Position: project → ./output/, one-time → $VISUAL/output/image-text-replace/

Baseline Task ID

See references/task-id-baseline.md corresponding line in.