DeepSeek Harness Plugin
中文

modlens

Featured
Tools & Capabilities 4061 stars DSH Plugin

The first vision plugin for DeepSeek Harness, and a vision bridge for text-only coding agents. Paste an image and receive structured JSON evidence back — OCR, layout, and semantic understanding — ready for any agent to act on.

Install this DSH plugin View source on GitHub

Plugin facts

GitHub stars 4061
LicenseMIT
Primary languageTypeScript
Last push2026-09-27T20:45:40Z
Maintainerliustack
Added to this directory2026-08-14

Screenshots

Install this DSH plugin

Copy the command below and run it in your DeepSeek Harness terminal.

dsh plugin add github:liustack/modlens

Plugins run with the permissions of your dsh process and may execute code at install time. Review the source repository and its license before installing. Pin the commit hash for reproducible installs. Plugin safety →

More about this plugin

🥇 The most capable vision plugin for DeepSeek Harness (dsh): install it instantly with one command: npx -y @deepseek-ai/dsh plugin --profile web add @liustack/[email protected]. See the setup guide for installation and update details. If the command line is not your thing but you still want to try DSH, check out AIManager , the lightest desktop wrapper for DeepSeek Harness. It gets you started with zero code or configuration and installs every dependency for you with one click.

Pasting an image works two ways. ① Just paste. On a text-only model the pasted image lands as a private temp file and its path enters the composer (the same interaction OpenCode and Pi ship), then the modlens read image tool takes it from there. ② Pick a (modlens vision) entry in the model selector (it remembers your choice, so once is enough), then paste: the thumbnail stays visible in your message, closer to the Codex app feel, and the image is converted to structured evidence at request time, answered by the same underlying route. The plugin auto-discovers every provider route carrying eligible text-only DeepSeek, GLM, or MiMo Pro models and adds a wrapped entry per route. A stock install gets DeepSeek-V4-Flash (modlens vision) and DeepSeek-V4-Pro (modlens vision) , while extra routes like opencode-go or zai get their own. Native vision models in those families, including GLM-5.3-Flash, are…

Use cases

modlens solves one problem: text-only models can't see, and you need them to see.

Debugging with screenshots. Paste an error screenshot or UI glitch into the conversation and modlens returns structured JSON — OCR text, layout regions, semantic description — that the model can reason about. You go from "describe the screenshot to me" to "here's exactly what's on screen."

Document extraction. Long PDF pages, slides, and screenshots become structured data the agent can query instead of guessing from alt text.

Frontend work. Show the agent a design mock or a rendered page; it gets the layout semantics it needs to write or fix frontend code accurately.

Zero-config start. The one-command install works without Python, and it routes to a compatible vision model transparently. If your workflow touches images more than twice a week, modlens pays for itself immediately. Pair it with dsh-vision-toolkit when you need long-screenshot OCR at scale.

Related guides

Learn how to choose, install, and use tools & capabilities plugins.

About this plugin: modlens

What is modlens?

modlens is a tools & capabilities plugin maintained by liustack. The first vision plugin for DeepSeek Harness, and a vision bridge for text-only coding agents. Paste an image and receive structured JSON evidence back — OCR, layout, and semantic understanding — ready for any agent to act on.

How do I install modlens?

Run dsh plugin add github:liustack/modlens in your DeepSeek Harness terminal. The dsh CLI resolves the plugin from GitHub and installs it into your active profile. For reproducible installs, pin a commit hash: dsh plugin add github:liustack/modlens#commit.

Is modlens free?

modlens is a community open-source project released under the MIT license. You can inspect its source and install it for free.