Which clients does it work with?

Claude and any client that supports remote MCP servers over Streamable HTTP.

Do I need an API key?

No. You sign in with OAuth when you connect.

Can an agent set up scheduled screenshots?

Yes. It can create, update, and delete them, and read what changed.

Yes, the MCP server is on every plan, including the free plan.

SCREENSHOT MCP INTEGRATION

Take website screenshots with AI agents, over MCP.

Connect Claude or any MCP client and your agent can capture pages, schedule watches, and read back what changed.

GET 25 FREE SCREENSHOTS → MCP guide

EXAMPLE AGENT TASK 4 steps

1

“Watch their pricing page daily” You PROMPT
2

create_screenshot_config PagePixels TOOL
3

list_all_change_notifications PagePixels TOOL
4

“The Pro plan went from $29 to $34” Agent REPLY

One prompt sets up a recurring watch that keeps running after the chat ends.

Screenshot MCP server for Claude and AI agents

The PagePixels MCP server lets AI assistants like Claude take website screenshots, which most AI chat tools can’t do on their own. Connect the server URL in any MCP client and your agent can capture a page, render HTML into an image, take screenshots from real locations, and analyze each image with AI to answer questions about live web pages.

The server exposes the full PagePixels API as native tools over Streamable HTTP with OAuth 2.1. That means an agent can also create scheduled screenshots, manage them later, and read change notifications, so a watch it sets up keeps running after the conversation ends. The MCP server is free to use on every PagePixels plan.

TOOLS THE AGENT CAN USE

The whole API, as native tools.

INSTANT

Capture now

Take a screenshot of a page or your own HTML right away.

snap · snap_html

SET UP

Running in four steps.

Create a free account

The MCP server is on every plan.
25 screenshots / mo
Add the server URL

Paste it into Claude or your MCP client.
mcp.pagepixels.com/mcp
Authorize

Sign in to PagePixels when asked.
OAuth 2.1
Ask for a screenshot

“Capture their pricing page.”
tool use

BROWSER AUTOMATIONS

Log in first. Then capture.

Fill fields, click, and wait for the right element before the agent captures it. Add a Real Location to see the page as a local visitor would.

01 Open the login page url
02 Fill email and password text_field ×2
03 Click submit click
04 Wait for the dashboard wait_for_selector
05 Capture the page png · full page
06 Return the image to the agent tool output

WHAT TEAMS BUILD

Jobs worth handing to an agent.

Competitive research
Watch a rival while you sleep

Set up a daily watch and ask what changed.
Reporting
Turn data into something visual

Render data as HTML and send the image to Slack.
Market checks
Same page, five countries

Capture one page from five Real Locations.
Research
Facts from many sites

Have AI pull the same details from a list of pages.

INCLUDED AT EVERY PRICE

Every PagePixels feature, available to your agent.

Free plan through 200,000 / mo

CAPTURE Browser automations Scheduled screenshots Real Locations in 150+ countries Ad, tracker, and banner removal CSS and JS injection

UNDERSTAND AI analysis Custom HTML screenshots HTML extraction

WATCH Change notifications Archived history CDN embed URLs

DELIVER Every integration + MCP server Full REST API with SDKs Custom headers and cookies

Which clients does it work with?

Claude and any client that supports remote MCP servers over Streamable HTTP.
Do I need an API key?

No. You sign in with OAuth when you connect.
Can an agent set up scheduled screenshots?

Yes. It can create, update, and delete them, and read what changed.
Is it free?

Yes, the MCP server is on every plan, including the free plan.