# luca-medeiros/lang-segment-anything

SAM with text prompt

Repository: https://github.com/luca-medeiros/lang-segment-anything
Canonical: https://ross.abutalabs.com/products/lang-segment-anything
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2025-08-28T00:12:25+00:00

## Health v2 (maintenance only)
Score: 42/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 39, release rhythm 21, longevity 88
- inputs: {"age_days": 1245, "days_push": 371, "days_rel": 371, "gap_med": 159.5, "n_releases_24m": 3}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2598, forks 300 (observed 2026-08-28T04:07:03.245043+00:00)

## What it is
A Python library combining Meta's Segment Anything Model 2 with GroundingDINO to generate segmentation masks for objects in images specified via free-form text prompts. It supports zero-shot text-to-bbox detection, batch inference, and deployment as an API endpoint.

## Use cases
- segment specific objects in an image using a text prompt
- zero-shot object detection with bounding boxes from text descriptions
- generate masks for objects like 'wheel' or 'person' in photos
- batch image segmentation with custom text prompts
- deploy a text-prompted segmentation model as an API endpoint
- annotate images for computer vision datasets

## When to choose
- you need to segment objects by name without training a custom model
- you want zero-shot detection and segmentation in one pipeline
- you need batch inference or a ready-made serving endpoint
- you want a simple Python API on top of SAM 2 and GroundingDINO

## When to avoid
- you need real-time segmentation on CPU-constrained hardware
- you require fine-grained class labels beyond text-prompt grounding
- you cannot download large model checkpoints or lack a GPU
- you need a lightweight pure-detection model without masks

## Facets
- artifact type: library
- maturity: active
- function: computer-vision, image-processing, machine-learning, deep-learning
- domain: computer-vision, image-processing, artificial-intelligence, machine-learning
- platform: python, cross-platform
- tags: segment-anything, grounding-dino, zero-shot-detection, instance-segmentation, text-prompt, sam2, object-detection, gradio, docker, gpu

## Member repositories
- luca-medeiros/lang-segment-anything (main) score 42

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:03.245043+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:21:20.933084+00:00, confidence not recorded.
  - readme: https://github.com/luca-medeiros/lang-segment-anything (fetched 2026-08-28T04:07:03.245043+00:00, sha 28e1b95ed1c5)
- Data as of 2026-08-30T08:39:29.467469+00:00.