dsh-tool-markitdown
Run the following command in DeepSeek Harness:
dsh plugin install jiavenzhong/dsh-tool-markitdown
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install jiavenzhong/dsh-tool-markitdown in your terminal to install the plugin (source at https://github.com/jiavenzhong/dsh-tool-markitdown), then restart dsh web to start using it.
About this plugin
Large language models work best with plain text, yet most real-world documents arrive as PDF, Word, Excel, PPT, EPUB, or notebooks. dsh-tool-markitdown leverages Microsoft MarkItDown locally to convert all these formats into clean, structured Markdown before the model ever sees them, eliminating the format barrier at the source.
The plugin offers three trigger paths: a one-click button in the input toolbar that converts a file and inserts the Markdown draft, automatic detection of file-path lines when you drag workspace files into the input and hit send, and a model-side markitdown tool that converts local paths, remote URLs, or uploaded attachments on demand. Every conversion runs on your machine; original files are read-only and never altered. Output exceeding the configurable character limit is truncated with an attachmentId the model can use to retrieve the full text later. All settings — Python command, size caps, timeout, auto-convert toggle — are adjustable in the Web GUI with instant effect.
Ideal for DeepSeek Harness users who regularly need an AI to interpret documents, spreadsheets, slide decks, or Jupyter notebooks, especially those working in the dsh web GUI. Pure JavaScript, no build step required — install and go.
Use Cases
- Drag a PDF report into the chat for instant model interpretation
- Let the model read Excel data tables or PPT slide key points
- Convert a Jupyter Notebook to Markdown for code logic analysis
Best For
- Developers who regularly need AI to interpret non-text documents
- Knowledge workers handling multi-format files in dsh web
- Workflow users who want to unify document formats into Markdown before model input
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
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.