You are a professional watermark cleanup assistant. Tas...
- Prompt
- You are a professional watermark cleanup assistant. Task: remove only watermark artifacts, including watermark text/logo overlays, repeated watermark grid/pattern, and semi-transparent watermark residues, ghosting, and blend halos. Hard constraints: (1) Preserve normal/original content text exactly as much as possible (titles, labels, packaging text, UI text, signage, subtitles); (2) Do NOT remove non-watermark text; (3) Keep subject identity, composition, geometry, perspective, color balance, lighting, and material texture unchanged; (4) Do not add new text, symbols, logos, or decorative patterns; (5) Output exactly one cleaned image with the same framing and style as source. Additional user request: You are a professional logo/watermark cleanup and background isolation assistant. Remove logo-like overlays, watermark residues, and text-like artifacts while preserving the real subject. Keep composition, perspective, lighting, and colors natural. Output one clean image with transparent-friendly, artifact-free edges around the subject.
- Size
- 1024 × 1536
- Model
- nano-banana
- Generation type
- text2img
- Created at
- 2026-04-20 18:30:08
How to use this deep-imager gallery example
This page is a moderated public-gallery still from the deep-imager AI image generator. It shows a real text to image result so you can study the prompt, model path, and framing before you recreate it with Make Same or start a fresh text to image / image to image run in the workspace.
What this still demonstrates
The creator’s prompt emphasizes: You are a professional watermark cleanup assistant. Task: remove only watermark artifacts, including watermark text/logo overlays, repeated watermark grid/pattern, and semi-transparent watermark residues, ghosting, and b….
When you remix, keep the subject and lighting cues that already work, then change only one variable at a time (ratio, background, or style). That saves credits compared with rewriting the entire brief.
Recommended remix workflow
- Click Make Same to load the prompt into the deep-imager workspace.
- Lock the aspect ratio for your channel (9:16 Stories, 1:1 feed, 16:9 thumbs) before Generate.
- Run one text to image pass, then refine with image to image if you only need lighting or background edits.
- Keep client drafts Private. Enable Publicly visible only when the still is ready for gallery review.
Credits, policy, and reporting
Credit cost appears on the Generate button before you confirm. Failed runs follow the platform refund path. Public listings use automated safety checks plus human review; machine blocks are separate from gallery approval status. If this image breaks the rules, use Report this image or email support@flyingmob.com. Full UGC rules: Terms of Service.
Learn more: User guide How credits work Prompt checklist Make Same from gallery FAQ
Trademark notice: Third-party model or brand names on deep-imager describe optional access, style compatibility, or comparisons only. deep-imager is not affiliated with those trademark owners.