AI Blocks › Build and ship software
Image object cutter
Click a point or draw a box on a picture. Get the exact outline of that object, or of every object.
Image object cutter is a free AI skill for the job "Build and ship software". Give the instructions to an AI assistant that supports the required files and tools, then review its output.
Free · No account · MIT licence · Made by Orchestra Research · Thinking: low
Made by Orchestra Research. See the original on GitHub; we copied it as it is, with its licence.
What it does
The exact outline of any object in a picture, from one click or a drawn box. Or let it find every object and cut them all out at once. It uses Meta's Segment Anything Model, which works on new kinds of images with no extra training. You can guide it with points, boxes or an earlier mask. Three model sizes trade speed for accuracy. A fast model suits quick tools and the most accurate suits detailed work. It can also run on its own across a whole image and return many masks. It suits annotation tools, training data and medical or satellite images. You need Python, torch and the segment-anything package.
You give it
- An image
- A click point or a box around the object (optional)
- Python with torch and segment-anything
You get
- A mask outlining the chosen object
- Masks for all objects in the image
- Confidence scores for each mask
- Code you can adapt for your own tool
Use it when
- You need a clean cutout of one object from a photo.
- You are building labeled data to train a vision model.
- You want every object in an image separated automatically.
- You are working with medical or satellite images with no labels.
How to use it
- Add it to your AI once: download it and add the folder to Claude, ChatGPT or Codex as a skill.
- Ask for the job in your own words. Your AI opens the skill when it fits.
- It works best when your AI can run Python code.