# microsoft/OmniParser

A simple screen parsing tool towards pure vision based GUI agent

Repository: https://github.com/microsoft/OmniParser
Canonical: https://ross.abutalabs.com/products/omniparser
Language: Jupyter Notebook
License: CC-BY-4.0
License Family: other
Last push: 2026-07-20T01:09:31+00:00

## Health v2 (maintenance only)
Score: 60/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 93, release rhythm 23, longevity 50
- inputs: {"age_days": 712, "days_push": 45, "days_rel": 355, "gap_med": 105.5, "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 25310, forks 2223 (observed 2026-08-28T04:11:38.109325+00:00)

## What it is
OmniParser is a screen parsing tool from Microsoft that converts UI screenshots into structured, understandable elements to ground vision-language models for GUI automation. It includes detection models for interactable regions and icons, enabling pure vision-based computer-use agents.

## Use cases
- parse ui screenshots into structured elements
- build a computer use agent that controls a gui
- ground llm actions to screen regions
- detect clickable icons and elements in screenshots
- automate desktop apps with vision models
- create training data for gui agents

## When to choose
- you need vision-based ui element detection and grounding for screen automation
- you want to drive a gui agent with models like GPT-4V, Qwen-VL, or Claude computer use
- you need state-of-the-art screen grounding benchmarks performance

## When to avoid
- you need DOM/accessibility-tree based automation rather than pure vision
- you need a production-ready end-user automation product rather than models and notebooks
- you cannot run GPU inference for the detection models

## Facets
- artifact type: library
- maturity: active
- function: computer-vision, machine-learning, ocr, agent-framework, image-processing
- domain: computer-vision, artificial-intelligence, large-language-models
- platform: python, cross-platform
- tags: gui-parsing, screen-understanding, ui-element-detection, vision-language-models, computer-use-agents, screenshot-parsing, ai-agents, gpu

## Member repositories
- microsoft/OmniParser (main) score 60

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:38.109325+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 2026-08-29T16:56:04.331520+00:00, confidence not recorded.
  - readme: https://github.com/microsoft/OmniParser (fetched 2026-08-28T04:11:38.109325+00:00, sha 8d09a52699eb)
- Data as of 2026-08-30T08:39:29.467469+00:00.