---
title: "Promptable segmentation with SAM 3"
description: "Segment every object matching a concept (a short text phrase or a bounding box) using Meta's Segment Anything Model 3."
url: "https://www.pixeltable.com/notebooks/img-promptable-segmentation"
updated: "2026-09-25"
---

# Promptable segmentation with SAM 3

Segment every object matching a concept (a short text phrase or a bounding box) using Meta's Segment Anything Model 3.

- **Type:** Notebook
- **Group:** Images

## What it covers

- Segment a single object from a text prompt
- Segment many instances of a concept at once
- Segment a specific region with a bounding box
- Segment objects across the frames of a video
- Apply SAM 3 to three customer use cases: satellite imagery, medical scans, and solar inspection

## What it uses

- **Install:** `pip install pixeltable torch torchvision transformers`
- **Pixeltable calls:** `huggingface.()`, `vision.overlay_segmentation()`, `video.clip()`, `huggingface.sam3_for_video_segmentation()`, `video.make_video()`, `vision.bboxes_draw()`, `image.tile_iterator()`, `image.stitch_tiles()`, `pxt.create_table()`, `pxt.create_view()`

## Open it

- **Open in Colab:** https://colab.research.google.com/github/pixeltable/pixeltable/blob/release/docs/release/howto/cookbooks/images/img-promptable-segmentation.ipynb
- **Docs page:** https://docs.pixeltable.com/howto/cookbooks/images/img-promptable-segmentation
- **Source on GitHub:** https://github.com/pixeltable/pixeltable/blob/release/docs/release/howto/cookbooks/images/img-promptable-segmentation.ipynb