mcpField
D
DocuPipe

DocuPipe

Local
docupipe.aiProductivity1Updated 2d ago
RepositoryVisit website

About DocuPipe

DocuPipe is a document-processing service for MCP clients that extracts structured JSON from files or URLs and can save the results. It also supports schema reuse, classification, splitting, and workflow operations.

Key features

  1. 01Extract structured JSON from files or URLs
  2. 02Create and reuse JSON Schemas
  3. 03Classify and split documents
  4. 04Run saved workflows
  5. 05Convert documents to Excel
  6. 06Check document status and extraction results

Use cases

  1. 01Extract invoice numbers, dates, and line items from PDFs
  2. 02Process multiple documents and save JSON results
  3. 03Organize uploaded documents by class
  4. 04Convert documents into local Excel workbooks

MCP support

DocuPipe provides an MCP server for document upload, extraction, classification, and conversion tools, callable from MCP clients such as Claude Code, Codex, and Cursor.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "docupipe-docupipe": {
      "command": "npx",
      "args": [
        "-y",
        "@docupipe/mcp"
      ],
      "env": {
        "DOCUPIPE_API_KEY": "<DOCUPIPE_API_KEY>",
        "DOCUPIPE_BASE_URL": "<DOCUPIPE_BASE_URL>"
      }
    }
  }
}
Run commandnpx -y @docupipe/mcp
Environment variables to set
DOCUPIPE_API_KEYSecretYour DocuPipe API key, from the Developers card on the Settings page.
DOCUPIPE_BASE_URLOverrides the API address. Defaults to https://app.docupipe.ai.

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npm@docupipe/mcp0.1.0

Tools

16

Saved tool list · updated 1d ago

ParameterTypeRequiredDescription
urlstringNoPublic URL of the file to upload, instead of file_path.
save_tostringNoOptional path to write the extracted JSON to, relative to the working directory.
citationsbooleanNoCite every value to its place on the page (bounding box and confidence). Runs at standard effort, 3 credits/page.
file_pathstringNoLocal file to upload, relative to the working directory.
schema_idstringNoSaved schema to extract with (list_schemas).
guidelinesstringNoPlain-language extraction instructions. Without a schema they also steer the drafted schema.
document_idstringNoAn already uploaded document, instead of uploading again.
json_schemaobjectNoInline JSON Schema (type object, properties with a description each). Saved as a new schema first, free.
schema_namestringNoName for the schema created from json_schema.
effort_levelenumNoExtraction effort. 'standard' (default) is right for most documents; use 'high' only after standard missed or misread values.standard | high
upload_job_idstringNoThe upload job of document_id, when resuming.
standardization_idstringNoResume: an extraction already started. Requires standardization_job_id.
include_field_metadatabooleanNoAlso return per-field page numbers and, with citations, boxes and confidence.
standardization_job_idstringNoResume: the job of standardization_id.

Details

From the project README

DocuPipe MCP server

Use DocuPipe from Claude Code, Codex, Cursor and other MCP clients. Ask your agent to "pull the line items out of these invoices" and it uploads the files, extracts the fields you named and saves the JSON, in one tool call per file.

Setup

  1. Sign up and copy your API key from the Developers card on the Settings page.
  2. Add the server to your client, replacing your-api-key. Node.js 18 or newer is required.

Claude Code

claude mcp add docupipe --scope user -e DOCUPIPE_API_KEY=your-api-key -- npx -y @docupipe/mcp

Codex

codex mcp add docupipe --env DOCUPIPE_API_KEY=your-api-key -- npx -y @docupipe/mcp

Cursor: open this link to install, then replace your-api-key in Cursor's MCP settings:

cursor://anysphere.cursor-deeplink/mcp/install?name=docupipe&config=eyJjb21tYW5kIjoibnB4IiwiYXJncyI6WyIteSIsIkBkb2N1cGlwZS9tY3AiXSwiZW52Ijp7IkRPQ1VQSVBFX0FQSV9LRVkiOiJ5b3VyLWFwaS1rZXkifX0=

Or add it by hand. Cursor and other clients that use an mcp.json file:

{
  "mcpServers": {
    "docupipe": {
      "command": "npx",
      "args": ["-y", "@docupipe/mcp"],
      "env": { "DOCUPIPE_API_KEY": "your-api-key" }
    }
  }
}

Then ask, for example:

Use DocuPipe to extract the invoice number, date, vendor and line items from every PDF in ./invoices and save one JSON file per invoice.

Tools

ToolWhat it does
extractFile or URL in, structured JSON out: uploads, extracts with a saved schema, an inline JSON Schema or a schema DocuPipe drafts, waits, and can save the result
upload_documentUpload a local file or URL and wait for it to be read
get_documentStatus, page count and parsed text of a document
standardizeExtract from an already uploaded document
get_standardizationFetch an extraction result
create_schema, list_schemasSave a JSON Schema, or reuse one the account already has
list_classes, classifyClassify a document into the account's classes
split_documentSplit a file into sub-documents, with AI choosing where each one starts
list_workflows, run_workflowRun a saved workflow on a file
convert_to_excel, get_excel_conversionTurn a document into an .xlsx workbook without a schema, and save it locally
get_job, wait_for_jobsCheck or wait on any job

Each call returns within about 45 seconds. When DocuPipe needs longer, the tool returns status: "processing" with every id created so far and the exact call to resume with. Resuming never uploads or extracts a second time.

Credits

Tools use credits from your DocuPipe account. Upload is 1 credit per page for PDF, image, Word and HTML files (CSV, Excel, JSON, text and XML are free to upload). Extraction is 2 credits per page at standard effort and 4 at high effort. See credits and billing for every operation.

Configuration

Variable
DOCUPIPE_API_KEYRequired. Your API key.
DOCUPIPE_BASE_URLOptional. Overrides the API address. Defaults to https://app.docupipe.ai.

Relative file paths resolve against the directory the server was started in, which for most clients is your project folder. Local files up to 100 MB are supported; pass a URL for larger files. Results larger than about 40,000 characters are written to docupipe-results/ in that folder and the tool returns the path with a preview.

Links

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.