A lightweight MCP (Model Context Protocol) server that wraps GroupDocs.Parser Cloud API to provide document-parsing capabilities to AI agents, assistants, and development environments.
It enables LLMs to extract text, images, and barcodes from documents stored in GroupDocs Cloud storage, and provides shared cloud storage utilities (upload, download, list, exists, delete).
- Lightweight GroupDocs.Parser Cloud MCP server
- Document parsing via GroupDocs.Parser Cloud:
- Extract plain text
- Extract images
- Extract barcodes
- Supports 50+ document formats (PDF, Word, Excel, PowerPoint, emails, archives, images, and more)
- Cross-platform: Windows, macOS, Linux
This section explains how to configure and run the GroupDocs.Parser Cloud MCP server
git clone git@github.com:groupdocs-parser-cloud/groupdocs-parser-cloud-mcp.git
cd groupdocs-parser-cloud-mcpCreate a .env file with your settings.
You can either create it manually or copy the .env.example file.
CLIENT_ID=your-client-id
CLIENT_SECRET=your-client-secret
MCP_PORT=8000
Get credentials at:
https://dashboard.groupdocs.cloud/#/applications
./run.sh.\run.ps1run.batServer starts at:
http://localhost:8000/mcp
This section describes integration examples.
- Run the inspector:
npx @modelcontextprotocol/inspector
- In the browser:
- Select “streamable HTTP”
- Enter your MCP URL:
http://127.0.0.1:8000/mcp - Click Connect
- Create
.vscode/mcp.json:
{
"servers": {
"groupdocs-parser-mcp-local": {
"type": "http",
"url": "http://127.0.0.1:8000/mcp"
}
}
}- Restart VSCode if required.
Now the MCP server is available inside VSCode’s MCP-enabled environments.
- Open KiloCode Settings → MCP Servers
- Open the Installed tab
- Edit “Project MCP”
- Add the config:
{
"mcpServers": {
"groupdocs-parser-mcp-local": {
"type": "streamable-http",
"url": "http://127.0.0.1:8000/mcp"
}
}
}- Add a test file such as
info.pdf - Example prompt:
Extract text from
info.pdfusing Groupdocs.Parser MCP and briefly summarize the document.
To use this MCP server with Cursor, configure the mcp.json file in your Cursor settings.
- Open Cursor Settings at the right top corner of the window
- Navigate to Tools & MCP section
- Click Add Custom MCP
- Locate or create the
mcp.jsonconfiguration section and add the following configuration:
{
"mcpServers": {
"groupdocs-parser-mcp-local": {
"url": "http://127.0.0.1:8000/mcp"
}
}
}- In a Cursor chat, feel free to reference documents stored in GroupDocs.Cloud or ask to process local documents so they can be uploaded to GroupDocs.Cloud. For example, use prompts such as:
Extract text from [your-folder]\a-book.epub with GroupDocs.Parser MCP and summarize the information from the e-book.
If you update the codebase and the MCP server stops working due to missing dependencies (for example, after adding new packages to requirements.txt), you may need to fully reinitialize the local virtual environment.
To do this, use the init_mcp script for your platform:
./init_mcp.sh.\init_mcp.ps1init_mcp.batThe init_mcp script performs a full reset:
- Removes the existing
.venvdirectory - Creates a fresh virtual environment
- Installs all dependencies from
requirements.txt - Prepares the environment so
run_mcp.*can start the server normally
After reinitialization, simply start the server again:
./run_mcp.sh.\run_mcp.ps1run_mcp.batThis ensures your environment is always in sync with the current project requirements.
This project is licensed under MIT License.