About Screenshotline
Screenshotline is an MCP server for agents that need rendered web pages as images or clean Markdown. It removes common page overlays and renders JavaScript-driven content before extraction.
Key features
- 01Capture pages as images
- 02Extract page text as Markdown
- 03Render JavaScript-driven pages
- 04Remove banners, ads and popups
Use cases
- 01Inspect a page layout visually
- 02Extract text from dynamic websites
MCP support
Exposes `screenshot` and `read_page` tools to MCP clients. The MCP bridge connects to a Screenshotline rendering API, including a self-hosted instance.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://api.screenshotline.com/mcpclaude mcp add --transport http mamubeenkhan-screenshotline https://api.screenshotline.com/mcp{"mcpServers":{"mamubeenkhan-screenshotline":{"type":"http","url":"https://api.screenshotline.com/mcp"}}}Open in Cursor{"servers":{"mamubeenkhan-screenshotline":{"type":"http","url":"https://api.screenshotline.com/mcp"}}}Open in VS Code{"mcpServers":{"mamubeenkhan-screenshotline":{"type":"http","url":"https://api.screenshotline.com/mcp"}}}{"mcpServers":{"mamubeenkhan-screenshotline":{"serverUrl":"https://api.screenshotline.com/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
https://api.screenshotline.com/mcpAuth requiredTools
2| Parameter | Type | Required | Description |
|---|---|---|---|
| url | string | Yes | The page to capture. Must be http or https. |
| format | enum | No | Default png.png | jpeg | webp |
| selector | string | No | Capture only the element matching this CSS selector. |
| block_ads | boolean | No | Block ad networks and collapse the empty slots. Default true. |
| full_page | boolean | No | Capture the entire scrollable page rather than just the viewport. |
| block_chats | boolean | No | Remove live-chat widgets. Default true. |
| block_popups | boolean | No | Remove newsletter and donation interstitials. Default true. |
| color_scheme | enum | No | Render the page in light or dark mode.light | dark |
| viewport_width | number | No | Viewport width in pixels. Default 1280. |
| viewport_height | number | No | Viewport height in pixels. Default 800. |
| block_cookie_banners | boolean | No | Remove consent banners. Default true. |
Details
From the project READMEscreenshotline-mcp
An MCP server that gives an agent two tools:
screenshot— capture a web page as an image, for when the layout is the answer.read_page— render the page, then return its text as clean Markdown, for when the text is.
Pages are fully rendered first: JavaScript executed, consent banners removed,
lazy images loaded, ads, chat widgets and newsletter popups stripped. So
read_page works on sites that return nothing useful to a plain HTTP fetch.
A capture that comes back blank is reported as a warning, not narrated as
though it were the page.
Install
Get a free key at screenshotline.com/account — 500 renders a month, no card. Then, in Claude Code:
claude mcp add screenshotline -e SCREENSHOTLINE_ACCESS_KEY=sl_live_... -- npx -y screenshotline-mcp
Or in any MCP client's JSON config (Claude Desktop, Cursor, and others):
{
"mcpServers": {
"screenshotline": {
"command": "npx",
"args": ["-y", "screenshotline-mcp"],
"env": { "SCREENSHOTLINE_ACCESS_KEY": "sl_live_..." }
}
}
}
Self-hosted
The renderer is open source. Point the server at your own instance and no key is needed unless you configured one:
SCREENSHOTLINE_BASE_URL=http://localhost:3000 npx -y screenshotline-mcp
| Variable | Default | |
|---|---|---|
SCREENSHOTLINE_ACCESS_KEY | — | Your API key. Required for the hosted API. |
SCREENSHOTLINE_BASE_URL | https://api.screenshotline.com | Any Screenshotline instance. |
This package is only the MCP bridge — one file plus the MCP SDK. It does not download a browser; the rendering happens on the API it points at.
Docs: screenshotline.com/docs/extract · License: AGPL-3.0-or-later
Comments
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.