kitchensink4word

🚰 KitchenSink4Word Community Edition

Landing page · llms.txt (machine-readable capability manifest for agents and LLM crawlers)

Edit Word documents with your AI assistant: tracked changes, citations, templates and live edits in Word.

Edit real Word documents from Claude Code, Codex CLI, Copilot CLI or any other MCP client that runs local tools. KitchenSink4Word connects your assistant to .docx files on Windows, and can edit a document while it is open in Microsoft Word so you see the changes land. Files are processed on your computer; the only thing that leaves it is what your AI app sends to its own provider. The Community edition is free under the AGPL. The Business edition adds a Windows installer, a signed update channel, a license your company can approve and support.

Works on: Windows. File editing works without Microsoft Word; live editing and the Word-powered features need Word installed.

Install

Pick the route for your AI app. The commands go in PowerShell on Windows or a terminal on macOS and Linux, not into an AI chat. The package routes need Python 3.12 or newer.

Claude Desktop

Install uv, then quit and reopen Claude Desktop. Download the .mcpb file from KitchenSink4Word releases. In Claude Desktop open Settings, then Extensions, then Advanced settings, then Install extension, and choose the file. The bundle fetches the Python package the first time it starts, so the first launch needs a network connection. Restart your session and check that the tools show as connected.

Claude Code or Codex CLI

Install uv, then run the line for your app and restart your session:

claude mcp add word -s user -- uvx kitchensink4word

codex mcp add word -- uvx kitchensink4word

Any other local MCP client

Use uvx as the command and kitchensink4word as its argument, or install the package and use kitchensink4word as the server command:

pip install kitchensink4word

Then follow your client's guide for adding a local MCP server. Installing the package on its own does not connect it to an AI app.

Business edition

Compare the editions on the pricing page. Already purchased? Your Windows installer and download link are in your license portal.

What it can do

110 document tools, plus the pack on/off controls.

• Revise a report while it stays open in Word, with edits you can watch.
• Read, accept and reject tracked changes, and work through reviewer comments.
• Fill a template or assemble chapters into one manuscript.
• Build and format tables, lists, headings and page layouts.
• Insert citations, bibliographies, footnotes and cross-references.
• Find formatting, accessibility and structure problems before handoff.
• Compare documents, prepare mail merges and redact selected text.
• Import or export PDF through Word when it is installed.

What is available depends on the packs you enable and the applications installed. The full tool reference is below.

Business edition

Need a license your company can approve and a setup someone supports? The Business edition pairs these tools with a Windows installer, a signed update channel and support under the Business terms. Update checks tell you when a covered release is available; nothing installs on its own. Compare the options on the pricing page. The Community edition stays free under the AGPL, including business use that meets its terms.

Privacy Policy

The tools run on your computer, and KitchenSink4AI receives no documents and no usage data from them. Your AI app may send prompts, file contents and tool results to its own provider under that app's settings and terms. Installing downloads packages, and the Community version check contacts PyPI unless you disable it; those requests carry connection details such as your network address and never your documents. Cloud folders and backups follow their own settings. The Privacy Policy covers the product, purchases and support records.

Not affiliated with, endorsed by, or sponsored by Microsoft Corporation. Microsoft and Word are trademarks of the Microsoft group of companies.

What the 222 operations cover

The everyday core covers text and formatting, tables (including merge-aware column insert/delete and one-call bulk cell edits), footnotes and endnotes (full lifecycle plus footnote and endnote conversion), TOC and caption lists, headers, footers, and sections, images, bulleted and numbered lists, content controls and fields, threaded comments, and tracked changes (read, accept and reject by author, and writing edits as tracked changes). Beyond that the surface spans equations (LaTeX to Word math), native charts, document assembly (chapter files into one manuscript), Zotero library citations, publication style conversion (8 styles, beta), review-cycle analytics, workflow suites (mail merge, batch operations, redaction, compliance and accessibility audits and fixes, submission prep, front matter, diagnostics), and Word-COM-backed document compare, field refresh, PDF export and import, and open-clean validation on Windows.

What makes it different

• Merge-aware table column operations. modify_table_structure inserts and deletes columns correctly through horizontally and vertically merged cells (gridSpan shrinks, vMerge chains re-root). At the time of writing, no other public Word MCP has this.
• Bulk-first API. Editing 20 cells is ONE set_cells call with a payload, not 20 round-trips.
• Tracked-change writing. track=True, author="Jane" on edit tools produces real Word revisions the recipient can accept or reject, proven round-trip against the server's own revision engine.
• Document compare. com_multi_document(action="compare") produces a Word-native redline between two versions of a document.
• Never corrupts. Atomic saves (temp file, structural validation, replace), automatic slot backups before every mutation, byte-identical passthrough of anything not being edited (equations, textboxes, content controls survive untouched), and clean typed errors: a file open in Word is refused with a message, not a hang.
• Fragmented-run safe. Find and replace works across Word's arbitrarily split runs while preserving per-character formatting, with a ReDoS timeout guard on user regex and an optional blast-radius limit.

Why this one (the capability comparison)

Every public Word MCP server was surveyed before building this. The honest comparison is about what each one can do, not how many names it has:

Capability	KitchenSink4Word	GongRzhe Office-Word (2.1k★, archived)	word-mcp-live (195★)	SecurityRonin docx-mcp (43★)
Tiered context loading (lite core, packs on demand)	✅ from ~9.9k tokens	❌	❌	❌
Live editing while the doc is open in Word	✅ cursor-safe, one Ctrl+Z per call	❌	✅	❌
Table column insert/delete	✅ merge-aware	❌	❌	❌
Bulk cell edits (one call)	✅	❌	❌	❌
Cell merge/unmerge	✅	merge only	❌	❌
Footnotes AND endnotes CRUD	✅ + conversion	add only	❌	✅
TOC insert + refresh	✅	❌	❌	❌
Native citations/bibliography	✅ 12 styles	❌	❌	❌
Index generation	✅	❌	❌	❌
Tracked-change WRITING	✅	❌	✅	✅
Accept/reject by author	✅	❌	partial	✅
Document compare + combine	✅ Word-native	❌	❌	buggy
Watermarks / protection / line numbers	✅	protect only	❌	❌
Section moving / template transfer	✅	❌	❌	❌
Atomic saves + auto-backup	✅	❌	❌	❌

Capability survey compiled from public repositories, documentation, and issue trackers. Corrections welcome: open an issue.

The packs

Numbers below come straight from scripts/measure_surface.py, never hand-counted, and every one counts the whole tools/list entry a client receives rather than the description and schema alone. The lite core loads at startup; the seven packs load on demand.

Pack	Tools	Approx tokens	What it carries
lite (startup)	29	~9.9k	Everyday reading and editing: text, paragraphs, tables, cells, lists, find and replace, outline, document view, backups, workflow guide, server info, pack toggles
references	8	~2.9k	Word-native citations and bibliography, Zotero search and cite, parity checks, style conversion and detection
review	9	~2.4k	Tracked changes (read, accept/reject, reports), threaded comments, structured diff, anonymize and deanonymize
academic	24	~7.6k	Footnotes and endnotes, TOC, index, captions, cross-references, front matter, chapter headers, sections, styles, list numbering, word counts, validation batteries, submission prep, accessibility
assembly	7	~2.1k	Insert and split documents, move sections, copy tables across files, apply and fill templates, mail merge
media-forms	16	~5.6k	Images, charts, equations, text boxes, hyperlinks, table structure and styling, form fields, content controls, field codes
com-live	13	~3.1k	Drives a local Microsoft Word: PDF import/export, compare and combine, proofing, readability, field refresh, live editing of open documents
protection-io	6	~1.6k	Document protection, watermarks, redaction with verification, table data import and export
Full surface	112	~35.2k	Everything (110 document tools plus enable_tools / disable_tools)

Quickstart: start lite, enable what you need

A session begins with the lite core. When a task needs more, the agent turns on the pack by name:

enable_tools(["references"])          # citations, bibliography, Zotero
enable_tools(["academic", "review"])  # notes, TOC, tracked changes, comments

Lite-tool refusals name the pack and the exact enable_tools call to run, and get_workflows recipes name the packs each workflow needs, so discovery is built in. To skip tiering entirely, start the server with KS4W_MODE=full and every tool is present from the first call.

Installed from the .mcpb bundle, the same two choices are checkboxes in the server's Claude Desktop settings: "Load every tool at startup" and "Lock the tool set at startup" (which fixes the surface so enable_tools refuses). They write KS4W_ALL_TOOLS and KS4W_LOCK_TOOLS. Both take true or false, treat empty as off, and refuse to start on anything else rather than guessing. KS4W_MODE beats the first and KS4W_PACK_POLICY beats the second whenever they are set, so a pinned pack list survives an install, and the server writes one line to stderr at startup naming what decided the surface.

One tip that removes most permission pop-ups: In Claude Desktop's Tool permissions for this server, set the Read-only tools group to Always Allow. Those tools cannot change anything, so approving them once is safe, and Claude stops asking about every read. Leave the acting tools on Ask, so anything that changes things still checks with you first.

One client-side caveat: some MCP clients drop a tool's schema when a pack is disabled and do not pick it back up on re-enable, even though the server sends tools/list_changed both ways (observed in Claude Code, 2026-09). If a re-enabled tool comes back as "no such tool", refresh the tool list on the client side (in Claude Code, a ToolSearch call for the tool reloads its schema).

Two numbers that matter

• 222 document operations, 110 tools. The operation count went up and the tool count came down on purpose. v1 spread similar jobs across many competing names; v2 gives each concept exactly one name built from a small verb table (insert_, set_, manage_, list_elements, validate, delete_element), so an agent picks the right tool the first time and carries less schema to do it. Fewer tools, more reach. Measured with one yardstick on both trees, v1.6 performed 200 operations across its 189 tools and v2.0 performs 222 across 110: every v1.6 capability survived the consolidation (the migration map covers all 189, test-guarded) and v2 adds the anchored batch editor, the anchored document view, deletion parity, and wider dispatch on the multiplexers. Both figures come from scripts/count_operations.py; the v1.6 run is scripts/count_operations_v16.py, which measures a v1.6 checkout with the same definition.
• Tiered loading: starts at about 9.9k tokens, scales to everything. A fresh session loads the 29-tool lite core (about 9,900 tokens) and turns on capability packs only when a task needs them, with one enable_tools call. Load every pack and the full surface measures about 35,200 tokens, so a session that never leaves lite carries roughly a quarter of the whole sink. (All figures are script-measured and count what the client actually receives; see Context cost below.)

Context cost (measured)

Almost no MCP server tells you what it costs to load. Here is the bill, from scripts/measure_surface.py:

• Lite start: 29 tools, about 9,900 tokens, loaded when the session opens.
• Full surface: 112 tools (110 document tools plus the two pack toggles), about 35,200 tokens with every pack enabled.
• Clients that defer tool schemas until first use (for example Claude Code) pay close to zero until a tool is actually called.

Both figures count the entire tools/list entry the client receives: name, title, description, input and output schemas, annotations and metadata. The estimator used to sum description and input schema only, which published a number about a quarter below what crosses the wire, and a server whose pitch is that it tells you what it costs does not get to publish the flattering subset. The v1.6 comparison this section used to carry was taken on that older yardstick, so it is gone rather than restated: no v1.6 measurement exists on this one.

Safety model

• Every mutating tool takes file_path and rotates the current content into stable backup slots before the change (backup=False to skip the rotation; the atomic validated save always applies). Backups live in a hidden .ks4w-backups/ folder next to the document, one subfolder per document, with exactly two slots: prev.docx (state before the most recent mutation) and anchor.docx (session start, rotating after 60+ minutes of idle). Storage stays bounded at roughly two copies per document no matter how many edits a session makes. manage_backups lists, restores (undoably: the pre-restore state rotates into prev first), and purges them, including leftover *.bak-* files from earlier schemes.
• Exclude .ks4w-backups/ from cloud sync tools (OneDrive, Dropbox, Google Drive): the slots churn on every edit and sync clients can hold locks that slow saves down.
• Mutations of the same file are serialized (in-process and across server processes via an advisory lockfile), so parallel calls cannot clobber each other and response metadata reflects settled document state.
• COM/live mutations (documents open in Word) remain outside this backup system; Word's own AutoRecover covers the open document.
• Saves are atomic and validated; a failed operation leaves the original byte-identical.
• Deleting content that carries footnote references automatically removes the now-orphaned definitions; validate reports integrity in both directions.
• Paragraph deletion refuses ranges that would cut a field (TOC, PAGEREF) in half or silently swallow a section break.

Update check (opt-out)

Variable	Effect
KS4W_UPDATE_CHECK	off turns the update check off completely: no network call, no cache file (the older KS4W_NO_UPDATE_CHECK=1 still works)

Update check. The server looks for a newer release on PyPI only when you call get_server_info, never at startup and never on a timer, at most one request every seven days, capped at two seconds. The check is a single plain HTTPS GET to pypi.org that sends nothing but the request itself. A failed check is reported with its reason rather than hidden. Set KS4W_UPDATE_CHECK=off to turn it off completely (the older KS4W_NO_UPDATE_CHECK=1 still works). The server never downloads or installs anything.

Sandboxing (opt-in)

Off by default: with nothing configured, the server behaves exactly as it always has. Set the KS4W_ALLOWED_ROOTS environment variable to a list of directories separated by the OS path separator (; on Windows, : elsewhere), for example %USERPROFILE%\Documents;D:\Work, and every path the server touches must resolve inside one of those directories. Reads are gated as well as writes, since a read outside the sandbox exfiltrates content just as surely as a write plants it. The containment check runs on canonicalized paths, so ..\ traversal, symlink and junction escapes, 8.3 short names, extended-length prefixes, case tricks, and lookalike sibling directories (Documents2 against an allowed Documents) are all caught, and UNC network paths are refused unless an allowed root is itself a UNC path that contains them. A blocked call refuses with a typed error naming the offending path and the allowed roots before any file is opened. Recommended whenever the server runs against untrusted or semi-trusted agent traffic.

Testing

2,068 tests (1,974 run everywhere; 94 live-marked tests drive a real Word instance on Windows): the suite was developed against a private corpus of real-world documents (book-length chapters, a document with 171 footnotes, a manuscript with 126 tracked changes and reviewer comments), and CI auto-generates structurally equivalent synthetic stand-ins (tests/make_corpus.py) so the full suite runs on any machine, including yours and every pull request. Local real documents, when present, take precedence. tests/word_validator.py opens outputs in invisible Word and fails on any repair prompt, the definitive corruption check.

Development history: prototyped with Claude Code in a day (2026-08-27), then hardened across release cycles through dedicated adversarial rounds (scale torture, pathological merge topologies, Unicode and schema fuzzing, ReDoS, Word-lock lifecycle, live-editing interaction hunts, COM leak checks) plus per-phase unit gates. Every finding fixed with a regression test, same session it was found. research/ documents the OOXML algorithms and pitfalls the implementation is built on, with attribution to the MIT-licensed reference implementations studied.

Maturity: what the version number does and does not claim

This project moves fast and is honest about what backs it. What the test record covers: every release passes the full suite plus dedicated adversarial rounds through the raw MCP transport, against a corpus of long, heavily formatted real-world documents with zero corruption across all of it. What it does not yet cover: other machines, Word builds older than current Microsoft 365, non-English Word installs (some tools reference styles by localized display name), RTL scripts, and the diversity of documents only real users bring. The safety net while the tool earns that mileage is structural: automatic slot backups before every mutation and atomic validated saves, so a bad outcome is a restore, not a loss. If something misbehaves on your documents, an issue with the symptom (never the document itself, unless it contains nothing private) is the most valuable thing you can send. Before filing: ask your AI to run get_server_info and paste the output here; it is designed to be safe to share.

Beta-labeled tools, heuristic by nature: review their flagged-items list rather than trusting silently: convert_citation_style, anonymize_for_review, validate(checks=["defined_terms"]), and the text-reference scan inside validate(checks=["cross_references"]). Each returns an explicit list of what it could not confidently handle.

Live-mode capability matrix

Which tools work on a document that is OPEN in Word. Dual-mode tools auto-route (live='auto'); everything else refuses with DOCUMENT_LOCKED until the file is closed. Every COM call is serialized server-side (one call reaches Word at a time, across separate server processes as well as threads in one), so concurrent agents queue instead of corrupting; live edits stay unsaved until com_save_document (the Option C model: see get_workflows(task='live-editing')).

Dual-mode tool (file + live)	Live-mode notes
apply_edits	one COM undo group; markdown lists/pipe tables in insert ops refuse up front with a close-the-file hint; replace verifies matched text in the live range; all other index-addressed ops run the snapshot staleness guard against the last saved state
delete_paragraphs	section-break/field-crossing ranges refuse; range={} location endpoints run the snapshot staleness guard
find_text	plain queries, 500-match cap; formatting mode and include_textboxes file-mode only
format_text	formatting mode routes live; case mode file-mode only (live='force' refuses); range={} locations run the snapshot staleness guard
get_comments	same shape
get_document_info	same key names; adds words/track_revisions, omits part list
get_outline	same flat-list shape
get_text	body shape identical; include_textboxes/textbox modes file-mode only
insert_paragraphs	heading_level maps to built-in Heading styles by numeric constant (outline-based docs get direct outline levels); inherit_format/copy_format_from file-mode only; text-selector locations run the snapshot staleness guard
search_and_replace	plain items route live (255-char finds handled); preview and find_formatting modes are file-mode only (find_formatting with live='force' refuses)
set_cells	plain text cells; vertical merges refuse live
set_paragraph_format	shading/borders/tab_stops keys refuse live (XML-level); raw indices keep the v1 index-trust contract
set_paragraph_text	expect guard honored; tracked-revision paragraphs refuse; text-selector locations run the snapshot staleness guard
word_count	Word's own statistics engine; exclusions mode file-mode only

Live-only tools (need the document open): live_insert_at_cursor, live_repair, live_scroll_to, live_set_track_changes.
COM tier (drive the Word application, open or closed): com_export_pdf, com_import_pdf, com_multi_document, com_proofing_errors, com_readability_statistics, com_refresh_fields, com_save_document, com_validate_opens_clean, com_word_status.
Open-document reader: get_document_view (reads the last SAVED state and says so).

The remaining 63 writers and 14 readers are file-only: they refuse while the document is open (close it, or com_save_document with close=true). The full per-tool table lives in docs/internal/live_parity_v2.md.

Known limits

• Live regex replacements skip matches positioned after complex fields in a story (COM character offsets drift there); the skip count is reported and a literal find still works. Live set_cells refuses vertically merged tables (the file-based tool is merge-aware; close the doc for those).
• Live tracked-change attribution is best-effort: Word signed into an Office account attributes revisions to that account; results report the effective author honestly.
• TOC and caption-list page numbers require a field update: automatic on next Word open, or immediate via com_refresh_fields.

Upgrading from v1.x

Every v1.x tool name changed. The 189-tool v1.6 surface was rebuilt as a consolidated set of 110 tools that cover every prior capability under one grammar. If you are upgrading from v1.x, read the migration guide first: it maps every old tool name to its v2 home, and get_workflows("migrate-from-v1") returns the same map in-session. New installs need nothing extra.

License

KitchenSink4Word is dual-licensed:

AGPL-3.0 (open source). Free for anyone (individuals, academics, and businesses) for any use that complies with the AGPL's terms. Those terms include sharing source, including your modifications, when you distribute the software or make it available over a network.

Commercial license. For organizations that want to build KitchenSink4Word into their own products or services without the AGPL's source-sharing obligations. Contact [e-mail].

Copyright (c) 2026 Alvut Consulting, LLC. KitchenSink4AI is a product line of Alvut Consulting, LLC.

Not affiliated with, endorsed by, or sponsored by Microsoft Corporation. Microsoft and Word are trademarks of the Microsoft group of companies.

Not affiliated with or endorsed by Microsoft Corporation. Microsoft and Word are trademarks of Microsoft Corporation. "For Microsoft Word" describes file-format compatibility only.