Skip to main content
Turn documents into reliable, reusable data. Extend helps you ingest PDFs, images, and scans, classify and extract key fields, and orchestrate document workflows so your teams can move from manual effort to automated, high-accuracy processes.
Built for document workflows, Extend lets you run processors, parse files, and manage runs with filtering and pagination so you can move from raw files to automated processing quickly.

​
How to Use MCP Nodes

​
What is Extend MCP?

Extend MCP gives AI-powered access to Extend so you can run processors, parse files, and manage workflow runs using natural language. You describe the action you want, the node calls the right Extend functions, and returns structured data ready for your workflow.

​
What Can It Do for You?

  • Orchestrate document workflows, trigger runs with files or text, then track statuses and outputs
  • Parse files into clean, chunked content for downstream processing or search
  • Run targeted processors for extraction, classification, or splitting on documents
  • Monitor and manage runs at scale with filtering, pagination, and lifecycle actions like cancel or delete

​
Available Tools

ToolWhat It DoesExample Use
List Workflow RunsList workflow runs with filtering by status, workflow ID, batch ID, and file name. Supports pagination and sorting.”List workflow runs filtered by workflow name and status status. Return run id, status, created at, and file name as structured data.”
Get Workflow RunGet details of a specific workflow run by ID to check status and output.”Using workflow run id, get the workflow run details and return status, started at, completed at, and output fields as structured data.”
Run WorkflowRun a workflow with files or raw text. Creates workflow runs for processing.”Run the workflow named workflow name with file at file url. Return workflow run id, status, and created at as structured data.”
Run ProcessorRun a processor on a document for extraction, classification, or splitting.”Run processor name on file at file url for extraction. Return processor run id, status, and processor type as structured data.”
List Processor RunsList processor runs with filtering and pagination.”List processor runs filtered by processor name and status status. Return run id, status, created at, and file name as structured data.”
Get Processor RunGet details of a specific processor run by ID.”Using processor run id, get processor run details and return status and output fields as structured data.”
Cancel Processor RunCancel a running processor run.”Using processor run id, cancel the processor run and return run id and final status as structured data.”
Delete Processor RunDelete a processor run permanently.”Using processor run id, delete the processor run and return confirmation and deleted run id as structured data.”
Parse FileParse files to get cleaned, chunked content in markdown or spatial formats. Supports sync and async modes.”Parse file at file url to markdown. Return chunks with text and page numbers as structured data.”
Get Parser RunGet status and results of a parser run.”Using parser run id, get the parser run status and return status and parsed content fields as structured data.”

​
How to Use

1

Create Your Extend MCP Node

Go to your node library, search for Extend, and click “Create a node with AI”
2

Add Your Prompt

Drag the Extend MCP node to your canvas and add your prompt in the text box. Use variables like file url, workflow name, or processor name so your node is reusable.
3

Test Your Node

Run the node to see the results. If it works as expected, you’re all set! If you run into issues, check the troubleshooting tips below.
4

Save and Reuse

Once your Extend MCP node is working, save it to your library. You can now use this customized node in any workflow.

​
Example Prompts

Here are some prompts that work well with Extend MCP: Workflow Orchestration: 
Run the workflow named `workflow name` with file at `file url`. Return workflow run id, status, and created at as structured data.
Run Processors for Extraction or Classification: 
Run `processor name` on file at `file url` for extraction. Return processor run id, status, and processor type as structured data.
Parse Files for Clean Content: 
Parse file at `file url` to markdown. Return chunks with text and page numbers as structured data.
Monitor and Retrieve Results: 
List processor runs filtered by `processor name` and status `status`. Return run id, status, created at, and file name as structured data.
Start simple with one focused action per node, include clear output fields, and use variables like file url, workflow name, and processor name. This keeps nodes fast, reliable, and ready for automation across many documents and workflows.

​
Troubleshooting

If your Extend MCP node is not working as expected, try these best practices:

​
Keep Prompts Simple and Specific

  • Good: “List workflow runs filtered by workflow name and return run id and status”
  • Bad: “Run a workflow on file url, keep checking until it completes, then parse the output and summarize it”
While this prompt might work, it is more efficient to break it into separate nodes. Extend MCP works best with focused, single-action prompts.

​
Match What Extend Can Do

  • Good: “Using processor run id, get processor run details and return status and output fields”
  • Bad: “Extract fields from a PDF, then write them to a Google Sheet and notify Slack”
Extend MCP focuses on document workflows. For sending messages or writing to spreadsheets, combine it with Gmail, Slack, or Google Sheets nodes in your workflow.

​
Break Complex Tasks Into Steps

For best results, avoid combining too many actions in one prompt since it can increase complexity and lead to timeouts: 
Parse all files from `folder url`, classify each document, extract key fields, and upload results to a spreadsheet with `sheet name`
Break this into smaller, focused nodes that each handle one task:
1

Step 1: Parse Files

Parse file at file url to markdown and return chunks with text and page numbers as structured data
2

Step 2: Classify Documents

Run classification processor name on file at file url and return processor run id and status as structured data
3

Step 3: Extract Fields

Run extraction processor name on file at file url and return extracted fields as structured data
In your workflow, connect these nodes sequentially. Map the output fields from the parse step into the classification step, then pass classification results into extraction.

​
Focus on Data Retrieval

Extend MCP is great at getting information from Extend. For analysis or content creation, connect it to other nodes. Example:
  • Good prompt: “Get details for processor run id and return status and extracted fields as structured data”
  • Bad prompt: “Get details for processor run id and analyze errors to recommend improvements”
Use the Ask AI node for analysis or summarization within the workflow. Extend MCP retrieves structured data from Extend so other nodes can analyze or generate content.

​
Troubleshooting Node Creation

If you’re seeing empty outputs in the node creation window (or if you’ve already created the node, hover over it and click “Edit”), use the chat interface to prompt the AI to add debug logs and verify the API response. Specifically mention that you received empty outputs.
In the node creation window (or if you’ve already created the node, hover over it and click “Edit”), use the chat interface to describe in detail what you expected versus what you received.
First click “Fix with Gummie”. If multiple attempts fail, simplify your prompt or contact support.
MCP node creation often requires a few tweaks. In the node creation window, use the chat or the “Request changes” button to refine filters, output fields, or pagination. The AI will adjust the node based on your feedback.

​
Need More Help?