# edititall-mcp [Health: Active]

**Category:** 🏢 Workplace & Productivity  
**Repository:** https://github.com/Subcue/edititall-mcp  
**GitHub Stars:** 0  
**npm Downloads (last month):** 159  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/edititall-mcp

## Description
Drive EditItAll's local in-browser editors: PDF, sheet, photo, vector, Word, slides.

## Tools
Capabilities this server exposes over MCP:

- **sheet_set_cells** — Write values or formulas (strings starting with "=") into the EditItAll spreadsheet. Cells use A1 notation. The sheet recalculates instantly.
- **sheet_read_range** — Read COMPUTED display values from the spreadsheet as a 2D array. Range in A1 notation, e.g. "A1:D10" or a single cell "B2".
- **sheet_load_csv** — Replace the spreadsheet with CSV content (bulk load — much faster than cell-by-cell). Provide inline `csv` text or a local file `path`. Formulas in cells (starting "=") are evaluated.
- **sheet_export_csv** — Export the spreadsheet's COMPUTED values as CSV — returned inline, and written to `path` if given.
- **pdf_open** — Open a local PDF in the EditItAll PDF editor. Returns the page count. (Best for files up to ~20 MB.)
- **pdf_export** — Export the PDF editor's current document (with all edits) to a local file.
- **pdf_rotate_page** — Rotate one page of the open PDF. page is 1-based; degrees is 90, -90 or 180.
- **pdf_delete_page** — Delete one page (1-based) of the open PDF. Returns the new page count.
- **pdf_merge** — Append the pages of another local PDF into the open document (after page `after`, default: at the end). Returns the new page count.
- **pdf_page_text** — Extract the text of one page (1-based) of the open PDF — lets you READ a PDF without uploading it anywhere.
- **word_write** — Type text into the EditItAll document editor at the caret (use \n for new paragraphs).
- **word_get_text** — Read the document editor's current content as plain text.
- **slides_from_outline** — Build a whole presentation in the EditItAll slides editor from an outline: a title slide plus one slide per entry (title + bullet list). Replaces the current deck.
- **vector_import_svg** — Import SVG markup into the EditItAll vector editor as a new document (the natural way for an AI to draw: generate SVG, then refine in the editor).
- **photo_open_image** — Open a local image file (png/jpg/webp) in the EditItAll photo editor as a new document.
- **photo_export** — Export the photo editor's current document (flattened) to a local file.
- **images_process** — Compress or convert local images through EditItAll's real codec pipeline. out: target format id (jpg|png|webp|avif|jxl|qoi) or "same" to recompress in place-format. Writes results next to the inputs (or into out_dir) and reports before/after sizes.
- **editor_screenshot** — Screenshot one of the editors so you can SEE the current state. app: sheet|vector|photo|tools|pdf|word|slides.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "edititall-mcp": {
    "command": "npx",
    "args": ["-y","@edititall/mcp"],
    "env": {
      "EDITITALL_URL": "",
      "CHROME_PATH": ""
    }
  }
}
```

**Requires environment variables:** `EDITITALL_URL`, `CHROME_PATH` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What edititall-mcp does

The edititall-mcp MCP server connects MCP clients to EditItAll, a local-first suite of browser-based editors for spreadsheets, PDFs, Word documents, presentations, vector graphics, photos, and image conversion or compression. It is intended for local file workflows where an agent needs to operate an editor rather than call a separate cloud document or media API.

Spreadsheet tools can write individual cells or formulas, load CSV data, export computed values, and read a range such as `A1:D10` as a two-dimensional array. PDF tools can open files, retrieve page text, rotate or delete pages, merge documents, and export the result. Document and presentation support covers writing or reading Word content and generating slides from a title and bullet outline.

Media tools open local images in the photo editor, export flattened results, import SVG markup into the vector editor, and process images through the codec pipeline. Image processing reports the input and output sizes. Screenshots can be requested for the sheet, vector, photo, tools, PDF, Word, or slides editor.

## How it works

The edititall-mcp MCP server starts a local headless Chrome, Edge, Chromium, or Brave process and loads EditItAll in a local browser profile. By default, the editor origin is `https://edititall.com`; development setups can point the server at a local `wrangler dev` URL. The server drives the editors’ official automation hooks and returns tool results to the MCP client.

Processing is performed on the device. According to the project documentation, the server does not upload files to Subcue AI LLC or a third-party API. The client sees only the results requested through tools, such as cell values, image-size changes, exported output, or screenshots.

## Setup and configuration

Node.js 22 or newer is required because the server uses Node’s built-in WebSocket support. A compatible local browser must also be installed. The npm package can be run with `npx -y @edititall/mcp`, installed globally, or started from the repository. A Rust wrapper is available through crates.io; it embeds the server module and executes Node, and its `NODE` setting can be used when Node is not on `PATH`.

Configure a stdio MCP client with `edititall-mcp` as the command, or use `node` with the path to the server module. Claude Desktop also has a downloadable extension bundle. Set `EDITITALL_URL` to change the editor origin and `CHROME_PATH` to select a specific browser executable; both have documented defaults.

## Tools and capabilities

- Use spreadsheet tools to write cells and formulas, load or export CSV, and read computed ranges.
- Use PDF tools to inspect page text, rotate or remove pages, merge files, and export PDFs.
- Use Word and slide tools to read or draft documents and create decks from outlines.
- Use vector and photo tools to import SVG, open images, and export flattened photos.
- Use image processing to convert or compress local images and compare sizes before and after.
- Request editor screenshots for supported editor surfaces.

## Limitations and notes

The edititall-mcp MCP server requires a locally installed compatible browser and Node.js 22 or later for the Node-based installation. The editor suite itself is closed-source, while this repository contains the MCP client/server that communicates with it. Browser selection and editor-origin overrides are available through environment variables.

_Full upstream README: https://allmcps.com/mcp/edititall-mcp/readme_