* [dsh-vision-router v2.0.0](https://github.com/ysr666/dsh-vision-router) – DSH Vision Router enables image-to-tool calling for text-only DeepSeek Harness agents using built-in vision chains and pixel-level tools without Python.

* [TongFlow v0.3.0](https://github.com/tong-io/tongflow) – Multi-modal GenAI workflow studio offering add, transform, and combine operations to orchestrate text, image, and speech generation into composed outputs.

* [doc7 v0.1.2](https://github.com/magicrew/doc7) – doc7 converts PDFs, Office files, scans, screenshots, charts, formulas, and diagrams into AI-ready Markdown using a local OpenAI-compatible multimodal model.