OpenScreenShot MCP
Web page screenshots for agents: a local MCP server and CLI driving your own Chrome, with secrets masked before pixels are written
- 93 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The server opens URLs in your browser and captures may contain personal data
- The image goes back to the client, which may pass it to a hosted model
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/openscreenshotDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The server opens URLs in your browser and captures may contain personal data
- The image goes back to the client, which may pass it to a hosted model
This is third-party code. Review the repository files before installing.
What it does
OpenScreenShot consists of a Chrome and Firefox extension and a separate optional `openscreenshot` package on npm with a CLI and an MCP server. The extension captures a whole page in one click, even with tricky scroll containers, and adds annotations, secret blurring, export to PNG, JPEG, WebP and PDF, and tab recording. Everything runs on the device with no account. The CLI and MCP are meant for agents. `npx openscreenshot shot https://example.com --out shot.png --full` captures a page in the Chrome, Chromium or Edge already on the machine. The MCP server starts with `npx -y openscreenshot` over stdio and exposes a `capture_screenshot` tool with arguments such as `url` and `fullPage`. The `--redact`, `--redact-pattern` and `--redact-rect` options cover secrets with black boxes before the image is written: by CSS selector, by pattern (email, phone, secret) or by rectangle. If a target moves, the capture is not saved. For clients that accept only remote MCP, there is a cloud option, OpenScreenShot Cloud, at `https://cloud.openscreenshot.app/mcp` with an API key.
Who it is for. Developers, DevOps engineers and testers who ask Claude Code to capture a page, dashboard or status page for reports and CI.
Good fit when
- You want a full-page capture attached to a ticket or an incident report
- An agent should see how the layout looks after an edit
- A pipeline needs screenshots for release notes and visual diffs
- You need email, phone numbers and tokens covered on the image before saving
- You want a quick capture with no account and no cloud service
Not a fit when
- There is no Chrome, Chromium or Edge on the machine: the CLI and server use an installed browser, and Node.js 22.12 or later is required
- Pages sit behind a login and you do not want to give the agent's browser access: the capture sees what the browser sees
- You need cloud keys for a client that cannot launch a local process: use Cloud
Example request
Capture https://staging.example.com/dashboard in full, cover email addresses and tokens on it and save it as dashboard.png.MCP
- Transport
- stdio
- Authentication
- not required
Limitations
Node.js 22.12 or later and an installed Chrome, Chromium or Edge are required. Cloud mode is optional: with `--cloud` or `cloud: true` the page URL and options go to the OpenScreenShot Cloud service, which needs a sign-in and a key. The tool returns a PNG to the client, which may send it to a hosted model, so use the masking options for sensitive data. Tab recording is available in Chrome only, and the project's desktop and web apps are still in development.
How to disable. Remove the `openscreenshot` entry from the MCP configuration with `claude mcp remove openscreenshot`. The extension is removed in the browser's extension manager.
FAQ
How do I connect OpenScreenShot to Claude Code?
With `claude mcp add openscreenshot -- npx -y openscreenshot`. The server runs locally over stdio, then you call the `capture_screenshot` tool.
Do I need the extension for MCP?
No. The CLI and MCP live in a separate `openscreenshot` package and drive your own Chrome, Chromium or Edge. The extension is for manual captures.
How do I hide secrets on the image?
With `--redact` and a CSS selector, `--redact-pattern` with email, phone, secret patterns and `--redact-rect` with a rectangle. If a target moves, the image is not written.
Does data go to a cloud?
Not by default; everything is local. Only with `--cloud` or `cloud: true` do the page URL and options go to OpenScreenShot Cloud.
Which formats are supported?
The extension exports PNG, JPEG, WebP and multi-page PDF and copies to the clipboard. The CLI example in the README saves PNG.
Discussion
No questions or reviews yet.
Related
MCP serversA skill helps an agent capture macOS windows for tests and documentation, choosing sizes, capture tools and output files
BrowserStack MCP Server: testing on real devices
BrowserStack MCP Server
The official BrowserStack MCP server: real devices, automated tests and test cases driven by an agent from your editor
The npx @wong2/mcp-cli command runs MCP servers from a Claude Desktop config, lists their tools and calls them from the terminal.
An MCP server and Chrome extension feed an agent console logs, network errors, exceptions, action recordings and page audits.
MCP Accessibility Scanner runs axe checks, exercises keyboard focus and saves accessibility reports for web pages
The sbx CLI and an MCP server run an agent's app in a clean Windows Sandbox and return a report with screenshots and crashes
README in short
The README introduces the extension with full-page capture, editing and export, then lists guides, install routes for Chrome and Firefox and a section on the CLI and MCP. Cloud mode, masking, DevOps and CI use, the MIT license and the roadmap are covered separately.