AI Agent Analysis

liustack/modlens

The first vision plugin for DeepSeek Harness, and the vision bridge for every text-only coding agent. Paste an image, get structured JSON evidence (OCR, layout, semantics). | 全网第一个 DeepSeek Harness 视觉插件，为 DeepSeek、GLM 等纯文本模型外挂视觉能力，粘贴图片即得结构化 JSON 证据（OCR、版面、语义）。

stars

1.9k

Language

TypeScript

GitHub

Source and compliance noteLast synced: Aug 16, 2026

Git-Stars is independent and not affiliated with GitHub or this project. Analysis may be AI-assisted and based on public repository metadata plus short README-derived summaries. We do not mirror full README files, docs, issues, or social comments.

Original GitHub source Methodology Editorial Policy

Overview

ModLens is a plug-in vision engine that gives text-only AI coding agents (like DeepSeek and GLM) the ability to read images pasted directly into the chat, converting them into structured JSON evidence (OCR, layout, semantics). It is the first vision plugin for DeepSeek Harness (dsh) and works across multiple harnesses with zero-config setup.

Installation

For DeepSeek Harness: `npx -y @deepseek-ai/dsh plugin --profile web add @liustack/modlens@3.16.7`; for skill-based harnesses, send the install line to your AI: 'Install and configure the modlens skill following https://github.com/liustack/modlens/blob/main/INSTALL.md, then run the health check and tell me the result.'

Problem solved

ModLens solves the problem of text-only models being unable to process visual information, which limits their usefulness in coding tasks that involve screenshots, diagrams, or UI mockups. It provides a lightweight, non-invasive bridge that adds vision capabilities without requiring model changes, proxies, or complex configuration, and it works across multiple popular coding harnesses.

What you can build

Developers can build agents that can analyze screenshots, extract text from images, understand UI layouts, and reason about visual content in coding workflows. Real-world use cases include automated UI testing, documentation generation from screenshots, debugging visual issues, and enhancing code review with visual context. The ceiling is high: with structured JSON evidence, agents can perform complex multimodal reasoning, such as comparing design mockups to implemented code or extracting data from charts and diagrams.

Community sentiment

Positive

No community feedback yet.

Concerns

No concerns documented yet.

Bottom line

ModLens is ideal for developers using text-only coding agents who need vision capabilities without heavy setup. It is perfect for those who want a quick, non-invasive way to handle images in coding workflows. However, teams with strict data privacy requirements should be cautious about sending images to external services, and those using models that already support vision may not need it. The key trade-off is convenience and breadth of support versus potential privacy and dependency concerns.

Analyzed by Git-Stars - 8/16/2026

liustack/modlens