Home/Plugins/Modlens

Modlens

Name: Modlens
Author: liustack

Add structured image evidence, OCR, layout, and semantic analysis to text-only agents.

GitHub stars★ 698

Maintainerliustack

LicenseMIT

StatusCommunity

About Modlens

Modlens is a vision bridge for DeepSeek Harness and other coding agents. In dsh it registers a native read_image workflow that turns pasted images into structured evidence, including OCR, layout, and semantic observations, before the text model answers.

Install in a dsh profile

$ npx -y @deepseek-ai/dsh plugin --profile web add @liustack/modlens@latest

Capabilities

◆Native read_image tool

◆OCR and layout evidence

◆Text-model vision bridge

◆Pasted image support

Developer preview note. DeepSeek Harness is changing quickly. Check the plugin README for its supported dsh version, prerequisites, and exact profile before installation.

Keep exploring

Related plugins

DSH Vision Toolkit

by Anionex

↗

Tools Skills145 stars

Give text-only agents eyes with image Q&A, OCR, grounding, UI restoration, and pixel diff tools.

npx -y @deepseek-ai/dsh plugin --profile web add github:Anionex/dsh-vision-toolkit

Custom Tool

by omdsh-dev

↗

Tools Sandboxes18 stars

Create, edit, and manage sandboxed JavaScript tools from the DeepSeek Harness web interface.

npx -y @deepseek-ai/dsh plugin --profile web add github:omdsh-dev/dsh-custom-tool

@File Mentions

by omdsh-dev

↗

Tools UI25 stars

Search and attach workspace files to prompts with fast, familiar Codex-style @file mentions.

npx -y @deepseek-ai/dsh plugin --profile web add github:omdsh-dev/dsh-at-file