Copilot Studio integration
Give your Copilot agent the ability to screenshot a website.
Add PagePixels Screenshots to the agent's toolset and it can capture a live page, turn HTML into an image, or shoot from another country, then hand the image to whatever comes next. Free plan included.
- Agents take web page screenshots, analyse them, and send the images to another tool.
- The custom HTML action turns HTML emails or data from other tools into images.
- The real geolocation action captures from different locations around the world.
- 20+ screenshot options, so the capture looks the way you asked for.
25 free screenshots / month · no credit card · no phone number
“Show me what the supplier’s returns policy page says today.”
-
Agent Chooses the screenshot tool tool use
-
Action Take a screenshot of a web page PagePixels
-
Reply Summary, with the dated capture attached image + text
How to add screenshots to a Copilot Studio agent
Four steps, and the last one is the only part that takes thought: telling the agent when to reach for a screenshot.
-
Create a free account
Sign up for PagePixels first, because the connector authenticates against an account. Every feature is on the free plan.
25 screenshots / mo -
Add the connector as a tool
In Copilot Studio, add PagePixels Screenshots to the agent's tools and connect it, the same way you would add any connector-based tool.
Tools → add -
Enable the actions you want
Give the agent the capture actions it should have: a plain page capture, custom HTML, a real-location capture, or an AI-analysed capture.
one to four actions -
Tell it when to use them
Describe the situations that call for a screenshot in the agent's instructions. That sentence is what turns an available tool into one the agent reaches for.
instructions
Our walkthrough is the Copilot Studio guide. Microsoft publishes the connector reference, action by action, at learn.microsoft.com, and the capture options themselves are in the API documentation.
Four actions, and 20+ options behind them.
These are the action names as Microsoft documents them, so what you enable in the agent matches what you read here.
-
Capture a web page
01The everyday one: the agent passes a URL and the options it wants, and gets the capture back to attach or pass on.
-
Take a screenshot of a web page
-
-
Capture your own HTML
02Render an HTML email or data the agent is holding into an image, for the things worth showing that were never a web page.
-
Take a screenshot of custom HTML
-
-
Capture from a real location
03Shoot the page as it appears in another country, city, or US state, so an agent can answer questions about other markets.
-
Take a real geolocation screenshot of a web page
-
-
Capture and analyse with AI
04Take the capture and get a written answer about it back with the image, so the agent has something to reason over as well as show.
-
Take a screenshot of a web page and analyze the image with AI
-
Behind the actions sit 20+ screenshot options: viewport size, full page, scale factor, format, element selectors, banner and ad removal, custom headers, and the rest. An agent can set them per call, so "the mobile view" and "just the pricing table" are things you can simply ask for.
An agent can describe a page. It cannot show you one.
Ask an agent what a page says and you get prose: confident, readable, and impossible to check without opening the page yourself. For anything that will be forwarded, filed, or disputed, the summary is the weakest part of the answer.
A screenshot turns the answer into evidence. The agent can still summarise, but now the message carries a dated picture of the page beside it, and the person reading can see for themselves in one glance.
It also lets an agent answer questions text alone cannot: how a page looks in another market, what an HTML email actually renders like, whether the layout is broken rather than just the wording changed.
- A summary you have to trust A dated picture you can check
- Cannot tell you how it looks Layout, banners, and pricing as rendered
- One view: the datacentre's The view from the market you asked about
- Nothing to forward or file An image that goes in the ticket or the thread
Building agents outside Microsoft 365, or want one that can schedule a watch and read back what changed weeks later? That is the MCP server, which exposes the whole API as native tools. This page is the Microsoft 365 route; they are not alternatives so much as different front doors.
The page your agent needs is often behind a login.
A supplier portal, an internal dashboard, a partner site that wants credentials. A renderer that only accepts a URL hands the agent a sign-in screen, and the agent will describe that instead, confidently.
Multi-step capture is what gets past it: log in, click links, fill out forms, run custom scripts, then capture. It is part of the capture engine on every plan, so the step list you build in the web app or the API is the same engine the connector calls.
Captures can also run from a Real Location in 150+ countries, states, and cities, which is one of the actions you can hand the agent directly.
-
Open the login page url
-
Fill email and password text_field ×2
-
Click submit click
-
Wait for the dashboard wait_for_selector
-
Capture the page png · full page
-
Hand the image to the agent connector output
Four agents that are better with a camera
In each case the agent decides a capture is needed. You only had to give it the tool.
-
support agent
See what the customer sees
A customer reports a broken checkout. The agent captures the page, attaches it to the ticket, and the person picking it up starts from a picture rather than a paraphrase.
web page capture · on request -
supplier and contract watch
Evidence for what a page said
Ask what a supplier's returns policy says today and get the answer with a dated capture beside it, which is the part that survives a dispute.
web page capture · dated archive -
marketing agent
Answer questions about other markets
Have the agent capture a campaign page from the countries it runs in, so questions about local pricing and consent screens get shown, not guessed.
real geolocation capture -
internal comms
Turn a rendered email into a picture
An HTML newsletter or report needs checking before it goes out. The agent renders it as an image so a person can approve what it will actually look like.
custom HTML capture
Agent access is not a separate tier.
This list is the same on the free plan and the largest one, and the same through the connector, the API, and the web app.
Capture
5 INCLUDED- Multi-step capture Log in, click, fill forms, and wait for elements before the shutter fires.
- Scheduled screenshots Recurring captures from every five minutes to once a year.
- Real Locations Capture through a proxy network in 150+ countries, US states, and cities.
- Ad, tracker, and banner removal Strip cookie consent walls, ads, and third-party scripts before capture.
- CSS and JS injection Restyle or manipulate the page before the capture is taken.
Understand
4 INCLUDED- AI analysis Ask a question of any capture and get a written answer back with the image.
- Custom HTML screenshots Render email, private spreadsheet rows, or API payloads into an image.
- HTML extraction Pull fully rendered HTML from JavaScript-built pages, whole page or one selector.
- Website domain research Extract structured fields across a list of domains, as JSON or CSV.
Watch
3 INCLUDED- Change notifications Watch a selector and get a webhook, Slack message, or Zap when it moves.
- Archived history Every capture is kept and listable through the API, with dated thumbnails.
- CDN embed URLs A permanent link that always shows the latest capture, no API key exposed.
Deliver
3 INCLUDED- Every integration Zapier, Make, n8n, Power Automate, Slack, Dropbox, webhooks, Chrome, and MCP.
- Full REST API Every option is a parameter, with NPM, PyPI, and RubyGem packages.
- Custom headers and cookies Send your own headers, cookies, user agent, language, and time zone.
Screenshots in Copilot Studio, answered.
Create a free PagePixels account, add PagePixels Screenshots to the agent's tools in Copilot Studio and connect it, enable the capture actions you want, and describe in the agent's instructions when a screenshot is appropriate. That last sentence matters more than the setup: it is what makes the agent reach for the tool.
Four, as Microsoft documents them: take a screenshot of a web page, take a screenshot of custom HTML, take a real geolocation screenshot, and take a screenshot and analyze the image with AI. Behind them sit 20+ screenshot options for size, format, full page, element selectors, and banner removal.
Yes. This integration, and every other one, is on every plan including the free one. There are no connector tiers and no per-agent fees. The free plan is 25 screenshots a month with no credit card and no phone number.
Same connector, different consumer. In Power Automate you decide the sequence and the flow runs it. In Copilot Studio you hand the agent the capability and it decides when a capture would help. Teams that do both usually keep scheduled reporting in a flow and leave ad-hoc questions to the agent.
When the agent is not built in Microsoft 365, or when it needs more than capturing: the MCP server exposes the whole API as native tools, so an agent can create a scheduled watch, list what is running, and read back what changed weeks later.
Yes, through multi-step capture, which is part of the capture engine on every plan: log in, click links, fill out forms, run custom scripts, then capture. Build the step list in the web app or the API and the connector calls the same engine, so an internal dashboard or supplier portal arrives as the real page.
