ClickRemoval
github.com
- Free plan
- Windows
- macOS
- Linux
- Web
- API
- Self-hosted
What ClickRemoval is
ClickRemoval is an open-source tool for removing objects from images and repairing the background without training, hand-drawn masks, or text prompts. You mark an object with positive clicks and protect areas with negative clicks. It supports Stable Diffusion 1.5, 2.1, SDXL, and compatible fine-tuned models. A Gradio web interface accepts an uploaded image and click placements, then returns a restored image. Command-line inference scripts support repeatable and batch-style processing, while the repository description lists a REST API. Docker and local installation are documented, with the demo available at localhost:7860. A shell script downloads supported model weights from author-maintained Hugging Face repositories. The README recommends an NVIDIA GPU with at least 24GB of VRAM, especially for SDXL at 1024 resolution. Lower-VRAM systems may use SD1.5 or SD2.1; SDXL Low VRAM mode reduces memory use but slows inference. ClickRemoval is free, Apache-2.0 licensed, and listed as a v1.0.0 pre-release.
Who it suits
It suits image editors and developers needing click-based object removal, restoration, batch inference, or API integration. Resource-constrained users can use the lighter SD1.5 or SD2.1 options.
Good to know, and what to watch
- Click-based removal without masks or text prompts
- Supports SD1.5, SD2.1, SDXL, and fine-tuned models
- Gradio web interface and command-line scripts
- Batch processing and REST API listed
- Apache-2.0 open-source license
- SDXL may require substantial GPU memory
- Low VRAM mode slows SDXL inference
- Version 1.0.0 is listed as a pre-release
- Model weights require separate downloads
The verdict
ClickRemoval focuses on interactive and scripted image repair with several Stable Diffusion backbones. Its hardware guidance and pre-release status matter when planning deployment.
What the maker publishes
- API
- The repository description says ClickRemoval has a REST API, though the README's quick-start instructions describe the Gradio demo and command-line scripts. source
- Audience
- The README presents SD1.5 as a lightweight option for resource-constrained devices and SDXL for higher-quality removal and restoration. source
- Click controls
- Positive clicks mark objects for removal, while negative clicks protect areas to keep. source
- Click interaction
- Users can mark objects to remove with positive clicks and protect areas with negative clicks, without masks or prompts. source
- Command line
- Inference scripts support reproducible testing and batch-style use. source
- Deployment
- The README documents Docker and local installation, with the Gradio demo opened at localhost:7860. source
- Diffusion models
- It supports SD1.5, SD2.1, SDXL, and compatible fine-tuned models. source
- Hardware
- The README recommends an NVIDIA GPU with at least 24GB VRAM for GPU inference, especially for SDXL at 1024 resolution. source
- Interaction
- The README describes ClickRemoval as training-free and says it removes objects by user clicks without hand-drawn masks or text descriptions. source
- Interfaces
- The project provides an interactive Gradio web demo and command-line inference scripts for batch-style use. source
- License
- The repository is distributed under the Apache-2.0 license. source
- Model downloads
- A shell script downloads SD1.5, SD2.1, SDXL, or all supported model weights from author-maintained Hugging Face repositories. source
- Models
- It supports Stable Diffusion 1.5, 2.1, and SDXL backbones, along with compatible fine-tuned models. source
- Purpose
- ClickRemoval is an open-source, training-free tool for click-driven object removal and background repair in images. source
- Release status
- The repository lists v1.0.0 as a pre-release. source
- Resource limits
- Lower-VRAM GPUs may run SD1.5 or SD2.1; SDXL may need Low VRAM mode, which reduces memory use but slows inference. source
- REST API
- The repository description lists a REST API among the tool's functions. source
- Security
- The README reports that an April 23, 2026 update fixed multiple bugs and vulnerabilities in inference code. source
- Support
- The project README provides installation and usage instructions and a Jupyter Notebook tutorial; it does not state a support service. source
- Web interface
- The project provides a Gradio demo that accepts an uploaded image and clicks, then returns a restored image. source
Checked against the maker's pages on 9 Oct 2026.