# Anionex/agent-vision-toolkit

为纯文本模型"看图“设计更好的视觉工具箱和技能，支持多图理解，图片问答，前端UI还原、GUI 自动化等，并可选无缝接入多个主流agent，直接识别粘贴图片｜ A vision toolkit and skill designed for text-only llms — image Q&A, long-screenshot OCR, frontend UI restoration, and GUI automation, with optional seamless integration for Codex, Claude Code, Pi, Oh My Pi, and OpenCode

Repository: https://github.com/Anionex/agent-vision-toolkit
Canonical: https://ross.abutalabs.com/products/agent-vision-toolkit
Homepage: https://agent-vision.anionex.me
Language: Python
License: MIT
License Family: permissive
Topics: codex, agent, deepseek, glm, multimodal, agent-skills, harness-engineering, claude-code, computer-use, opencode, vision, vision-language-model, text-only-llm, dsh-plugin
Last push: 2026-08-25T15:10:04+00:00

## Health v2 (maintenance only)
Score: 79/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 98, longevity 2
- inputs: {"age_days": 32, "days_push": 8, "days_rel": 19, "gap_med": 7, "n_releases_24m": 2}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1108, forks 38 (observed 2026-08-28T04:03:36.573852+00:00)

## What it is
A Python toolkit of vision CLI tools plus an agent skill that gives text-only LLM agents (like DeepSeek) image understanding capabilities, including image Q&A, long-screenshot OCR, frontend UI restoration, and GUI automation. It optionally integrates via a local proxy and plugins with agents such as Codex, Claude Code, Pi, and OpenCode.

## Use cases
- let a text-only LLM agent see and answer questions about images
- run OCR on long screenshots from a coding agent
- restore frontend UI from a screenshot
- automate GUI interactions with a text-only model
- paste images into a non-multimodal coding agent
- add vision tools to DeepSeek or other text-only model agents

## When to choose
- your agent runs on a text-only model and lacks multimodal support
- you use Codex, Claude Code, Pi, or OpenCode and want vision capabilities
- you need screenshot OCR or UI restoration inside an agent workflow

## When to avoid
- your model is already natively multimodal
- you need a standalone OCR service rather than agent tooling
- you don't use shell-capable coding agents

## Facets
- artifact type: library
- maturity: active
- function: computer-vision, ocr, agent-framework, cli, mcp, image-processing
- domain: artificial-intelligence, computer-vision, developer-tools, large-language-models
- platform: python, cli, cross-platform
- tags: vision-tools, text-only-llm, agent-skills, gui-automation, screenshot-ocr, ui-restoration, claude-code, codex, proxy-integration, ai-agents

## Member repositories
- Anionex/agent-vision-toolkit (main) score 79

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:36.573852+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 2026-08-30T06:43:43.120678+00:00, confidence not recorded.
  - readme: https://github.com/Anionex/agent-vision-toolkit (fetched 2026-08-28T04:03:36.573852+00:00, sha ba93bc91d90e)
- Data as of 2026-08-30T08:39:29.467469+00:00.