Inpaint Anything
Inpaint Anything is a free, open-source project for removing, filling, or replacing selected objects in images, videos, and 3D scenes.
What Inpaint Anything is
Inpaint Anything is a free, open-source project for removing, filling, or replacing selected objects in images, videos, and 3D scenes. In its image workflow, you click an object for segmentation and an inpainting model such as LaMa fills the gap. Fill Anything uses a text prompt to create content in a selected area, while Replace Anything uses a prompt to change the background. Video processing follows a selected object across frames and fills those areas with a video inpainting model such as STTN. For 3D work, Remove Anything 3D tracks an object across source views, inpaints those views, and uses NeRF to create new views without it. The project supports any image aspect ratio and 2K resolution. You can run its local web interface on your own machine or try the linked Hugging Face interface. Installation requires Python 3.8 or newer, PyTorch, model dependencies, and downloaded checkpoints.
Good
- Free and open-source
- Handles images, videos, and 3D scenes
- Text prompts guide fills and replacements
- Supports any aspect ratio and 2K images
- Local web interface available
Watch out
- Requires Python, PyTorch, dependencies, and checkpoints
- Documented setup requires downloaded model files
- Some models have non-commercial licenses
- Defaults target a single 24GB GPU
Who it suits
It suits users who want self-hosted, prompt-guided object editing across images, video, or 3D scenes. Windows users can follow the project’s Miniconda-based setup.
Verdict
Inpaint Anything combines segmentation, inpainting, prompts, and view tracking in one project. Its broad media support is useful, but setup, hardware, and model licensing require attention.
The facts
Each one read from Inpaint Anything's own pages, with the date it was checked.
What it does
- 3D processing
- Remove Anything 3D tracks a selected object across source views, inpaints those views, and uses NeRF to synthesize novel views without the object.checked 4 Oct 2026
- Beta branch
- The README describes the main_2026 branch as beta and says it requires Python 3.12 or newer, PyTorch 2.7 or newer, and CUDA 12.6 or newer.checked 4 Oct 2026
- Fill and replace
- Text prompts guide filling selected objects with new content or replacing backgrounds.checked 9 Oct 2026
- Hardware
- The repository says its defaults target a single 24GB GPU, and FLUX.1-Fill can require about 20GB of VRAM.checked 9 Oct 2026
- Hosted demo
- The README links to a Hugging Face interface for trying Inpaint Anything online.checked 4 Oct 2026
- Image capabilities
- The README lists support for any aspect ratio and 2K resolution.checked 4 Oct 2026
- Integrations
- The README identifies Segment Anything, LaMa, Stable Diffusion, OSTrack, and STTN as models used or acknowledged by the project.checked 4 Oct 2026
- License
- The repository includes an Apache License, Version 2.0, which grants no-charge copyright and patent permissions subject to its terms.checked 4 Oct 2026
- Local interface
- The repository says its demo can run as a local web UI; the main_2026 quick start launches it at localhost:7860.checked 9 Oct 2026
- Media support
- The repository describes workflows for images, videos, and 3D scenes, with support for any aspect ratio and 2K resolution.checked 9 Oct 2026
- Model downloads
- Users need to download model checkpoints and place them in the documented project directories to run the workflows.checked 4 Oct 2026
- Model licensing
- The README notes that ProPainter and FLUX.1-Fill-dev are non-commercially licensed and says to check the terms for LaMa, SDXL, and SAM 3 before commercial use.checked 9 Oct 2026
- Models
- The main_2026 branch uses SAM 3, LaMa, SDXL, and ProPainter by default in different stages, with other backends available.checked 9 Oct 2026
- Object removal
- Users can select an object by clicking it or, on the main_2026 branch, naming it with a text prompt, then fill the removed area.checked 9 Oct 2026
- Prompted editing
- Fill Anything uses a text prompt to generate content in a selected area, while Replace Anything uses a prompt to change the background.checked 4 Oct 2026
- Purpose
- Inpaint Anything uses segmentation and inpainting models to edit images, videos, and 3D scenes.checked 9 Oct 2026
- Requirements
- The main_2026 branch requires Python 3.12 or later, PyTorch 2.7 or later, and CUDA 12.6 or later for its default SAM 3 setup.checked 9 Oct 2026
- Research authors
- The project lists Tao Yu, Runseng Feng, Ruoyu Feng, Jinming Liu, Xin Jin, Wenjun Zeng, and Zhibo Chen as authors.checked 9 Oct 2026
- Robotics workflow
- The main_2026 branch includes batch hand removal from egocentric footage and mask export for Human-to-Robot pipelines.checked 9 Oct 2026
- Video processing
- Remove Anything Video tracks a selected object across video frames and uses a video inpainting model such as STTN to fill it.checked 4 Oct 2026
The company
- Founded
- 2023checked 28 Sep 2026
Where it runs
- Windows setup
- The original README gives a Windows installation path using Miniconda and a Windows-specific requirements file.checked 9 Oct 2026
Alternatives to Inpaint Anything
- 1
Recoll
Free planOpen sourceAndroidWebAPIExtension7.9Free plan - 2
GitHub CodeQL
Free planWebExtensionLinuxmacOS7.5From $30 a month - 3
Open Semantic Desktop Search
Free planOpen sourceWebAPIExtensionLinux7.5Free plan - 4
Agent Ransack
Free planWindows6.6$20 one-time - 5
Magpie
Free planWindows6.6Free plan - 6
WizFile
Free planWindows6.6$20 one-time