baizhichen-web / dsh-vision-fallback-bridge

Listed

Lightweight DeepSeek Harness upload bridge: images and common documents become workspace paths so text-only DeepSeek models can read them with Qwen-MM-Plugins vision tools.

mainModelTool View source

Installation

npx -y @deepseek-ai/dsh plugin --profile web add github:baizhichen-web/dsh-vision-fallback-bridge

This installation command is an unverified starting point generated from the GitHub repository address.

README

Maintainer-authored documentation snapshot.

View on GitHub ↗
Commit 81b09e3Synced Aug 18, 2026

dsh-vision-fallback-bridge

Lightweight DSH upload bridge for text-only DeepSeek models.

  • Images keep the native DSH attachment UI and are converted to workspace paths on send.
  • Common documents (pdf/docx/xlsx/pptx/txt/md/csv/json/zip) are queued as simple chips and converted to workspace paths on send.
  • The model receives type-labelled paths, so it can route images to Qwen-MM-Plugins vision tools and documents to file tools.

Install

Copy this directory into ~/.dsh/profiles/web/node_modules/@dsh-external/dsh-vision-fallback-bridge, then add @dsh-external/dsh-vision-fallback-bridge to dsh.profile.bundles in ~/.dsh/profiles/web/package.json.

Restart DSH Web profile.

Storage

Uploaded files are written to:

<workspace>/.dsh-vision/uploads/

Supported types

TypeExtensions
Imagespng, jpg, jpeg, webp, gif
Documentspdf, doc, docx, xls, xlsx, ppt, pptx, txt, md, csv, json, zip

License

MIT

Repository information

Language
JavaScript
License
Not reported
Last updated
Aug 16, 2026, 4:14 PM

Install deliberately

Review source code, permissions, lifecycle hooks, dependencies and network access. Test untrusted plugins in an isolated environment.