修復 text layout

Branch logic

• If the user has not provided image で bad text, ask 向け it before generating the final clean image.
• If a reference image is provided, preserve the subject identity, proportions, と important details before changing style または scene.
• If the first result fails ビジュアル inspection, fix the smallest broken area first instead of regenerating the whole image.
• If the first tool path fails, switch to the next listed option: GPT Image 2, Recraft.
• If the task depends on masks, text, または object edges, inspect those details at full size before accepting the 出力.
• If the image will be 公開済み commercially または publicly, run the risk guardrails before export.

When not to use

• Do not use this as proof that the generated clean image is factually real または legally approved.
• Do not use it when the user needs professional legal, medical, safety, または regulated-compliance judgment.
• Do not publish the result until 権利, consent, authenticity, と platform constraints have been checked.

• The final asset clearly matches the requested clean image.
• The main subject still reflects the original image で bad text.
• Text, hands, logos, edges, と 商品ディテール are not visibly broken.
• The result is usable in the target channel 変更せずに another full regeneration.
• Proofread all 表示中 text, labels, logos, と small typography before using the final image.