Name: DocuPipe
Author: docupipe

DocuPipe MCP

Upload documents and extract structured JSON from Claude Code, Codex, Cursor and other MCP clients.

Grade A Local · stdio API key 16 tools v0.1.0 MIT

No reviews yet by docupipenpm Runs on your machine, nothing hosted

96/100

trust score

tools declared

local

runs on your machine

—

downloads a day

What it does

Upload documents and extract structured JSON from Claude Code, Codex, Cursor and other MCP clients.

Quickstart

# 1 — add it to Claude Code
claude mcp add --scope user --env 'DOCUPIPE_API_KEY=${DOCUPIPE_API_KEY}' --transport stdio docupipe-mcp '--' npx -y @docupipe/mcp

# 2 — ask your agent for something these tools do
#   extract, upload_document, get_document and 13 more

It needs DOCUPIPE_API_KEY: this reads it from your environment, so export it first.

DocuPipe costs nothing on mcprush: there is no plan to choose here, no cap to set and nothing mcprush can bill you for.

Publishernot claimed

docupipe1 server

1 category

All 1 server

Collected from a public index. Nobody has claimed this account, so nothing here was written by its author — claim it if it is yours.

Installation

Where are you running it?

Every route below installs the same thing and ends at the same approval screen. This one runs on your machine: your client starts DocuPipe as a process under your own user, with your files and your network, so the tool surface below is what it can do to your computer rather than to a server somewhere else. Its scan result is on the Security tab — read the surface before you approve it.

Claude Codedeclared by this listing

This is a public server: you run it yourself and this marketplace is not in the path. You need Node.js — the current LTS (24) is the safe choice; npx comes with it. Claude Code registers it in one command; --scope user makes it available in every project. Claude Code reads DOCUPIPE_API_KEY (Your DocuPipe API key, from the Developers card on the Settings page) from your environment each time it starts the server, so the value stays out of its config: export it in your shell profile (macOS, Linux), or run setx DOCUPIPE_API_KEY … and open a new terminal (Windows). The '--' is in quotes so that PowerShell passes it on.

claude mcp add --scope user --env 'DOCUPIPE_API_KEY=${DOCUPIPE_API_KEY}' --transport stdio docupipe-mcp '--' npx -y @docupipe/mcp

Reconnect, or start a new session, and the tools appear in the model’s tool list.

Free on mcprush, and it runs from its publisher’s own package: this marketplace is not in the path, so there is no key of ours to hold and no card to add.

What it asks of you

It reaches a system of yours, so it needs your own credential rather than the publisher’s: DOCUPIPE_API_KEY — Your DocuPipe API key, from the Developers card on the Settings page.. You set it in your own client’s config, on your machine — this marketplace never holds it.

What it writes

One config entry your client uses to start the process locally. A local server runs with your file system and your network, which is why it is priced without metering.

What you approve

16 tools, with what each one reads, writes and reaches shown before you agree — the same list on every route above. Read the tool surface.

Tool surface

What the model actually sees. Descriptions are diffed on every release — see version history.

Only tools that write or reach the network

Read from its source code. A server that builds its tool list when it starts can list more tools than this.

One call from a file to structured JSON: uploads a local file or URL (or reuses a document_id), runs a DocuPipe extraction, waits and

Parameter	Type	Required	Notes
file_path	string	no	Local file to upload, relative to the working directory.
url	string	no	Public URL of the file to upload, instead of file_path.
document_id	string	no	An already uploaded document, instead of uploading again.
upload_job_id	string	no	The upload job of document_id, when resuming.
schema_id	string	no	Saved schema to extract with (list_schemas).
json_schema	object	no	Inline JSON Schema (type object, properties with a description each). Saved as a new schema first, free.
schema_name	string	no	Name for the schema created from json_schema.
guidelines	string	no	Plain-language extraction instructions. Without a schema they also steer the drafted schema.
effort_level	string	no	Extraction effort. 'standard' (default) is right for most documents; use 'high' only after standard missed or misread values.
citations	boolean	no	Cite every value to its place on the page (bounding box and confidence). Runs at standard effort, 3 credits/page.
standardization_id	string	no	Resume: an extraction already started. Requires standardization_job_id.
standardization_job_id	string	no	Resume: the job of standardization_id.
save_to	string	no	Optional path to write the extracted JSON to, relative to the working directory.
include_field_metadata	boolean	no	Also return per-field page numbers and, with citations, boxes and confidence.

Upload a local file or URL to DocuPipe and wait for it to be read (OCR and parsing). Returns the document_id used by every other tool.

Parameter	Type	Required	Notes
file_path	string	no	Local file, relative to the working directory.
url	string	no	Public URL of the file, instead of file_path.
dataset	string	no	Optional dataset name to group documents.
metadata	object	no	Optional metadata stored with the document.
wait	boolean	no	Wait for parsing to finish (default true).

Fetch a document's processing status, page count and parsed text (tables as markdown). Set include_layout for per-page

Parameter	Type	Required	Notes
document_id	string	yes	—
include_layout	boolean	no	Return the full per-page layout (sections, tables, boxes) instead of plain text.

Run an extraction (standardization) on one uploaded, fully processed document and wait for the result. Use extract instead when

Parameter	Type	Required	Notes
document_id	string	yes	—
schema_id	string	no	Saved schema; omit to have DocuPipe draft one from the document.
guidelines	string	no	—
effort_level	string	no	Extraction effort. 'standard' (default) is right for most documents; use 'high' only after standard missed or misread values.
citations	boolean	no	Cite every value to its place on the page (bounding box and confidence). Runs at standard effort, 3 credits/page.
wait	boolean	no	Wait for the result (default true).
save_to	string	no	Optional path to write the extracted JSON to, relative to the working directory.

Fetch the extracted JSON of a finished standardization. While the extraction is still running this answers not found; wait on its

Parameter	Type	Required	Notes
standardization_id	string	yes	—
save_to	string	no	Optional path to write the extracted JSON to, relative to the working directory.
include_field_metadata	boolean	no	Also return per-field page numbers and, with citations, boxes and confidence.

Save a JSON Schema as a DocuPipe schema so every extraction returns the same fields. Free. Use type object with a description on

Parameter	Type	Required	Notes
schema_name	string	yes	—
json_schema	object	yes	—
guidelines	string	no	Plain-language instructions applied whenever this schema is used.

List the account's saved schemas (id and name), newest first, to reuse one instead of creating a new one. Free.

Parameter	Type	Required	Notes
limit	integer	no	—
offset	integer	no	—

List the account's classification classes (id, name, description). classify needs class ids from here. Free.

No parameter schema on file.

Classify one uploaded document into the account's classes and wait for the answer.

Parameter	Type	Required	Notes
document_id	string	yes	—
class_ids	array	no	Candidate classes (list_classes). Omit to use all classes.
multi_class	boolean	no	Allow more than one class per document.
include_unknown	boolean	no	Allow 'unknown' when nothing fits (single-class mode).
instructions	string	no	—

Split one uploaded document into sub-documents, with AI deciding where each one starts, and wait for the new document ids.

Parameter	Type	Required	Notes
document_id	string	yes	—
instructions	string	no	How to split, e.g. 'one document per invoice'. Can also skip parts you don't need.
dataset	string	no	Dataset for the new documents.

List workflows that run_workflow can start on one document: classic workflows and Architect workflows that are runnable and take a

No parameter schema on file.

Upload a local file or URL with a workflow attached and, for Architect workflows, wait for the run's outputs. Classic workflows return

Parameter	Type	Required	Notes
workflow_id	string	no	Workflow to run (list_workflows).
file_path	string	no	—
url	string	no	—
dataset	string	no	—
metadata	object	no	—
workflow_run_job_id	string	no	Resume: the run job returned earlier.

Turn a document into an .xlsx workbook without a schema (one sheet per table). Takes a local file, URL or document_id, waits for

Parameter	Type	Required	Notes
file_path	string	no	—
url	string	no	—
document_id	string	no	—
upload_job_id	string	no	The upload job of document_id, when resuming.
instructions	string	no	What the spreadsheet should contain, e.g. which tables or columns to keep.
schema_id	string	no	A schemaId from an earlier conversion, to lay out similar files the same way.
save_to	string	no	Optional local path for the .xlsx file.

Check an Excel conversion, wait for it to finish, and optionally save the .xlsx locally. The download link is fresh on every call. Free.

Parameter	Type	Required	Notes
conversion_id	string	yes	—
save_to	string	no	Optional local path for the .xlsx file.

Fetch one job's status, type, error message and results (for split, classify and workflow jobs). A job in error is returned as a

Parameter	Type	Required	Notes
job_id	string	yes	—

Wait until the listed jobs finish (any job type) and report each one's status. Returns early when all are done or one fails; call

Parameter	Type	Required	Notes
job_ids	array	yes	—
max_wait_seconds	integer	no	—

Free

$0 forever

Every tool, no call limit
No card, and mcprush charges nothing for it
Source published under MIT
Runs on your machine — nothing of it reaches our gateway
Nothing to cap, because mcprush bills nothing

Nothing here is billable: mcprush charges nothing to install DocuPipe or to call it, at any volume.

ATrust score 96 / 100

Release 0.1.0 · scanned 7 Nov 2026

Two independent axes, because powerful and malicious are different questions. The grade is threat only. The capability level is blast radius, and it is never a penalty on the grade — it is priced as one subtract-only term in the score, where you can see it.

Threat score100

Client score96

Findings0 threat · 5 capability

Capabilitymoderate

Coveragesource

Verificationrepo

Surface digestb33636dd59e8efe6f1d8f1fe52bae151e825d1de645b22ad73aeef0f4f16ba92

Score vectorthreat 100 → client 96

Term	Level	What it prices	Points
capability-exposure	moderate	capability blast radius (moderate) — client exposure if the model is manipulated	−3
verification-discount	repo	publisher verification (public source) — no provenance, but the source is public and inspectable	−1
coverage-honesty	source	inspection depth (source) — how much of the target the scan could see	−0

score = clamp(round(threatScore − E_cap − E_ver − E_cov), 0, 100). No model, no opinion, no network call at scoring time: the same methodology version against the same bytes always produces this number, and every line above is a field of the report rather than a summary of it.

What the scan could actually read

A grade is only as meaningful as its coverage, so the scanner publishes its own depth before it publishes its result.

Tool surfaceread

Implementation sourceread

Package metadataread

Live transportnot available

Tools were statically extracted from the published source (16 recovered), not enumerated from a running server. Tool-poisoning, Unicode-smuggling, capability and toxic-flow analysis ran on this inferred surface, but a mis-parsed registration could be missed or mis-attributed, so tool-derived findings are capped below “confirmed”. To grade the real runtime surface, scan the running server: --command "npx -y <package>".

Capability — what it could do if the model were manipulated

Tags derived from each tool’s schema and the implementation, not from what the tool calls itself. moderate is the level these add up to.

Tool	Capability tags	Why the tag was assigned
extract	untrusted-inputexternal-sink	untrusted-input: declares openWorldHint: true (interacts with entities outside the server); external-sink: handler sends an HTTP POST ( client.post(path) at dist/tools.js:103
upload_document	untrusted-inputexternal-sink	untrusted-input: declares openWorldHint: true (interacts with entities outside the server); external-sink: handler sends an HTTP POST ( client.post(path) at dist/tools.js:103
get_document	untrusted-input	untrusted-input: keyword "fetch", declares openWorldHint: true (interacts with entities outside the server)
standardize	untrusted-inputexternal-sink	untrusted-input: declares openWorldHint: true (interacts with entities outside the server); external-sink: handler sends an HTTP POST ( client.post(path) at dist/tools.js:103
get_standardization	untrusted-input	untrusted-input: keyword "fetch", declares openWorldHint: true (interacts with entities outside the server)
create_schema	untrusted-inputexternal-sink	untrusted-input: declares openWorldHint: true (interacts with entities outside the server); external-sink: handler sends an HTTP POST ( client.post(path) at dist/tools.js:103
list_schemas	untrusted-input	untrusted-input: declares openWorldHint: true (interacts with entities outside the server)
list_classes	untrusted-input	untrusted-input: declares openWorldHint: true (interacts with entities outside the server)
classify	untrusted-inputexternal-sink	untrusted-input: declares openWorldHint: true (interacts with entities outside the server); external-sink: handler sends an HTTP POST ( client.post(path) at dist/tools.js:103
split_document	untrusted-inputexternal-sink	untrusted-input: declares openWorldHint: true (interacts with entities outside the server); external-sink: handler sends an HTTP POST ( client.post(path) at dist/tools.js:103
list_workflows	untrusted-input	untrusted-input: declares openWorldHint: true (interacts with entities outside the server)
run_workflow	untrusted-inputexternal-sink	untrusted-input: declares openWorldHint: true (interacts with entities outside the server); external-sink: handler sends an HTTP POST ( client.post(path) at dist/tools.js:103
convert_to_excel	untrusted-inputexternal-sink	untrusted-input: declares openWorldHint: true (interacts with entities outside the server); external-sink: handler sends an HTTP POST ( client.post(path) at dist/tools.js:103
get_excel_conversion	untrusted-input	untrusted-input: keyword "download", declares openWorldHint: true (interacts with entities outside the server)
get_job	untrusted-input	untrusted-input: keyword "fetch", declares openWorldHint: true (interacts with entities outside the server)
wait_for_jobs	untrusted-input	untrusted-input: declares openWorldHint: true (interacts with entities outside the server)

Toxic-flow graph

The lethal trifecta, checked as a graph rather than as a checklist: untrusted input, a sensitive source and an external sink have to meet before there is a path worth worrying about.

flow-1

Untrusted input can drive an external action even though no sensitive source is exposed.

Supply chain and provenance

Integrityfirst-seen

Surface digestb33636dd59e8efe6

Scanned2026-11-07

This is the first scan of this surface here, so there is nothing yet to compare it against.

Every result on this tab comes from one deterministic pass over the published package — offline, rule by rule, and auditable line by line above. Same methodology version, same bytes, same score.