ComfyUI Node
Auto Watermark Remover
A ComfyUI node in image/cleanup with 27 inputs and 4 outputs.
Auto Watermark Remover
- image
- watermark_template
- symbol_reference
- image
- mask
- mask_preview
- detected_text
◄languagesen►
◄detection_modecv2_only►
◄regionfull_image►
◄min_confidence0.25►
◄padding10►
◄dilate5►
◄blur3►
◄cv2_sensitivity0.80►
◄corner_fallbackoff►
◄corner_width_ratio0.12►
◄corner_height_ratio0.08►
◄inpaint_methodopencv_telea►
◄inpaint_modelbig-lama.pt►
◄inpaint_radius5►
◄gpufalse►
◄detail_detectornone►
◄symbol_match_threshold0.45►
◄diffusers_model_idstable-diffusion-v1-5/stable-diffusion-inpainting►
◄diffusers_promptclean image, preserve the original subject and style, remove watermark and text►
◄diffusers_negative_promptwatermark, text, logo, signature, artifact, distortion, deformed anatomy, blurry face►
◄diffusers_strength0.90►
◄diffusers_guidance_scale4.5►
◄diffusers_steps30►
◄diffusers_crop_padding48►
Categoryimage/cleanup
Inputs (27)
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE | — | |
| languages | STRING | en | Used only when OCR modes are selected. |
| detection_mode | COMBO | cv2_only | How to create the watermark mask before cleanup. cv2_only is usually the best starting point for logos and symbols. |
| region | COMBO | full_image | Limits detection to likely watermark areas. Smaller regions are faster and reduce false hits. |
| min_confidence | FLOAT | 0.250–1 | Minimum OCR confidence before text is masked. Only matters for OCR modes. |
| padding | INT | 100–256 | Extra pixels around each detected area before cleanup. |
| dilate | INT | 50–256 | Expands the final mask. Increase if the watermark edges are still visible after removal. |
| blur | INT | 30–255 | Softens the mask edge. Lower values keep the repaired region tighter. |
| cv2_sensitivity | FLOAT | 0.800–1 | Sensitivity for non-OCR detection. Raise it for faint overlays, lower it to avoid masking real image detail. |
| corner_fallback | COMBO | off | Adds a simple mask in one or more corners when the watermark sits near the border and detection is inconsistent. |
| corner_width_ratio | FLOAT | 0.120.02–0.5 | Corner fallback width as a fraction of image width. |
| corner_height_ratio | FLOAT | 0.080.02–0.5 | Corner fallback height as a fraction of image height. |
| inpaint_method | COMBO | opencv_telea | Cleanup backend. Start with opencv_telea for small marks, big_lama for thicker texture overlap, symbol_reverse_blend when you have a watermark_template for a semi-transparent symbol. |
| inpaint_model | COMBO | big-lama.pt | Big-LaMa checkpoint name. Used only when inpaint_method=big_lama. |
| inpaint_radius | INT | 51–64 | OpenCV or cleanup radius. Small values preserve detail better; larger values can erase thicker marks. |
| gpu | BOOLEAN | false | Use GPU for OCR when available. Big-LaMa and Diffusers manage their own device placement. |
| detail_detector | COMBO | none | Extra detector for hard masks. Use symbol_template together with watermark_template for known logos or repeated symbol watermarks. |
| symbol_match_threshold | FLOAT | 0.450–1 | Template match threshold for symbol_template and symbol_reverse_blend. Raise it to be stricter. |
| diffusers_model_id | STRING | stable-diffusion-v1-5/stable-diffusion-inpainting | Optional Hugging Face model id or local folder for diffusers_sd_inpaint. |
| diffusers_prompt | STRING | clean image, preserve the original subject and style, remove watermark and text | Prompt used only by diffusers_sd_inpaint. Keep it focused on preserving the original subject while removing the watermark. |
| diffusers_negative_prompt | STRING | watermark, text, logo, signature, artifact, distortion, deformed anatomy, blurry face | Negative prompt used only by diffusers_sd_inpaint. |
| diffusers_strength | FLOAT | 0.900–1 | How strongly Diffusers redraws the cropped area. Lower values preserve more of the source image. |
| diffusers_guidance_scale | FLOAT | 4.51–20 | Prompt guidance for diffusers_sd_inpaint. |
| diffusers_steps | INT | 301–100 | Inference steps for diffusers_sd_inpaint. |
| diffusers_crop_padding | INT | 480–512 | Extra context around the masked area when using diffusers_sd_inpaint. |
| watermark_templateopt | IMAGE | Optional reference image of the watermark to match. Use a tight crop when possible. A clean external template works better than a noisy crop from the same image. | |
| symbol_referenceopt | IMAGE | Legacy alias for watermark_template. You only need one of these inputs connected. |
Outputs (4)
| Name | Type | Description |
|---|---|---|
| image | IMAGE | — |
| mask | MASK | — |
| mask_preview | IMAGE | — |
| detected_text | STRING | — |