mimo-vision
Run the following command in DeepSeek Harness:
dsh plugin install wulusai2333/mimo-vision
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install wulusai2333/mimo-vision in your terminal to install the plugin; source is available at https://github.com/wulusai2333/mimo-vision
About this plugin
When your main model has no vision input, asking the AI to understand an image usually means juggling extra toolchains or workarounds. mimo-vision exists to close that gap: as a native DeepSeek Harness plugin, it registers a single describe_image tool that sends an image to a mimo-v2.5 multimodal model and returns a plain-text description to the main model-a lightweight vision bridge with nothing else in between.
Under the hood, the plugin natively handles PNG, JPEG, GIF, WebP, and BMP. For extended formats like SVG, TIFF, HEIC, PSD, AVIF, and JXL, it automatically transcodes them to PNG via a local ImageMagick install and downscales to a 2048px long edge to keep token usage lean. Routing is practical: a free tier is tried first, with a one-shot paid fallback on failure, and a single flag to disable the paid path entirely. Architecturally, registration, dependency declaration, and teardown all ride on DSH native capability seams, so a clean uninstall is a structural guarantee rather than a cleanup script.
It is best suited for developers working inside DeepSeek Harness whose main model lacks vision but who regularly need the AI to describe or analyze images-design mockups, screenshots, user uploads, and the like. One natural-language sentence is all it takes to invoke the tool; no separate multimodal inference pipeline required.
Use Cases
- Give a non-vision main model the ability to describe images
- Batch-process SVG, HEIC, TIFF and other extended formats into text descriptions
- Let the AI analyze design mockups, screenshots, or user uploads
Best For
- Developers on DSH whose main model lacks vision input
- Teams needing a lightweight multimodal inference bridge
- Solo devs who want to skip building their own inference pipeline
Related Plugins
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.
An armor-breaking plugin for DeepSeek: rewrites instruction priority, expands refusal-blocked phrases, persists across sessions, and shows a green active indicator.