Skip to main content

​
Install and Configure the Plugin

1

Install the plugin

Log in to your Dify platform, navigate to Tools → Plugin Marketplace, search for SoMark, and add the plugin.
2

Configure plugin settings

After installation, open the plugin configuration page:
  • API Base URL: Defaults to https://somark.tech/api/v1. No changes needed in most cases.
  • API Key: Enter your SoMark API Key.
No API key yet? Go to the SoMark API workbench to get your API key. Free quota does not need to be claimed and is automatically credited to your account (500 pages/day, 2000 pages/month).
3

Save

Click Save — the plugin is now ready to use in workflows.

​
Using SoMark in a Workflow

1

Add the SoMark tool node

In the Dify workflow editor, click + to add a new node, choose Tool, then find and add the SoMark > SoMark Document Parsing node.Add SoMark tool node in Dify
2

Configure input variables

Click the variable icon {x} in the input field and select the file variable provided by an upstream node (e.g. sys.files from your Start node). You can also configure optional parameters (for example Output Formats, Image Format, Table Format) as needed. If you leave Output Formats empty, the node returns both Markdown and JSON by default. See Input Parameters below for details. The API Key is handled automatically by the plugin configuration — no need to enter it in the node.Configure input variable in Dify
3

Reference outputs in downstream nodes

Once the node executes, its output variables are available to all downstream nodes (LLM, text splitter, code node, etc.). Click {x} in any downstream node’s input field to select the SoMark node’s output variables.

​
Parameters and Outputs

​
Input Parameters

ParameterTypeRequiredDescription
Filefile✅Supported files: PDF, PNG, JPG, JPEG, BMP, TIFF, JP2, DIB, PPM, PGM, PBM, GIF, HEIC, HEIF, WEBP, XPM, TGA, DDS, XBM, DOC, DOCX, PPT, PPTX. Max 200 MB / 300 pages.
Output Formatscheckbox❌Select one or more output formats. Supported options: Markdown, JSON. If left empty, the default outputs are Markdown and JSON.
Image Formatselect❌Image output format. Supported options: URL, Base64, None. Default: URL.
Formula Formatselect❌Formula output format. Supported options: LaTeX, MathML, ASCII. Default: LaTeX.
Table Formatselect❌Table output format. Supported options: HTML, Markdown, Image. Default: HTML. In Markdown mode, merged cells are expanded into individual cells with duplicated content.
Chemical Structure Formula Formatselect❌Chemical structure output format. Supported options: Image. Default: Image.
Enable Text Cross Pageboolean❌Merge text that spans across pages into a continuous paragraph. Default: False.
Enable Table Cross Pageboolean❌Merge tables that span across pages into a continuous table. Default: False.
Enable Title Level Recognitionboolean❌Recognize heading hierarchy such as H1/H2/H3. Default: False.
Enable Inline Imageboolean❌Return images embedded in text paragraphs. Default: True.
Enable Table Imageboolean❌Return images embedded in table cells. Default: True.
Enable Image Understandingboolean❌Perform semantic understanding and structured description for images in the document. Default: True.
Keep Header Footerboolean❌Keep page headers and footers instead of filtering them out. Default: False.

​
Output Variables

VariableDescription
markdownParsed document content in Markdown format, preserving the original layout including headings, tables, lists, equations, and images
json_strThe complete raw JSON response string returned by the SoMark API, containing detailed structured data for each document element (type, content, bounding box, page number, etc.). Ideal for advanced processing in a code node