ModLens is a vision plugin for DeepSeek Harness that extends text-only coding models with image understanding capabilities.
The tool solves the problem that DeepSeek's flagship chat models and GLM are text-only and cannot process images. ModLens works by accepting images pasted directly into chat and returning structured JSON evidence including OCR results, layout analysis, and semantic information. This approach eliminates the need to save images to files and pass file paths manually.
Developers should choose ModLens if they work with DeepSeek Harness or other text-only language models and need vision capabilities without switching to multimodal models. The tool installs as a single command within the DeepSeek Harness ecosystem. For those preferring a graphical interface over command-line setup, the README mentions AIManager as a desktop wrapper that handles installation and dependency management automatically.
The maintainers respond to issues and pull requests within hours. Work in the issue tracker centers on bug reports, enhancement requests, and user questions.
