Ask about and manage your sites and uptime monitors with your AI assistant
Connect AI platforms and tools such as Claude, ChatGPT, Codex, or VS Code to WebPixie over MCP. The assistant reads your monitoring data and analysis findings, answers in plain language, and helps you decide what to do next. Sign in with your WebPixie account, choose what the assistant can see, and decide whether it can change anything.
Connect your assistant
Add this address to your assistant as an MCP server or custom connector, then sign in with WebPixie.
https://app.webpixie.io/mcpSetup guides
See what an assistant can answer
Four example conversations. The first two work with read-only access.
acme.io went down this morning. What happened?
Your monitor for https://acme.io timed out at 09:14.
Three checks in a row got no response within 30 seconds.
That points to the server or a service in front of it. Check its load and restart logs around 09:14.
What did WebPixie’s analysis find on acme.io, and what should I fix first?
2 findings to review on acme.io
High · Missing Strict-Transport-Security header
Medium · Missing Content-Security-Policy header
Start with the High finding.
Add shop.acme.io to WebPixie with SSL and DNS monitoring.
Added shop.acme.io.
SSL and DNS monitoring are on.
Pause the staging monitor while we deploy.
Paused “staging-api”.
It stays paused until you resume it.
Example conversations are illustrative, and answers depend on your data. Adding a site or pausing a monitor needs read and write access.
What the assistant can work with
Reading works in every workspace you connect. Write actions need read and write access and follow your workspace role.
Uptime monitors
Can read: Status, check history, and uptime summaries.
With write access: Create, update, pause, resume, or delete monitors.
Sites
Can read: Site details and which checks are turned on.
With write access: Add, update, or remove sites.
SSL and certificates
Can read: Certificate details, expiry dates, and the certificates seen for a domain.
Domain and DNS
Can read: Registration and expiry details and DNS records.
Links and sitemaps
Can read: Crawled links with their status, and the pages listed in your sitemap.
Incidents
Can read: Open and past incidents and the check details behind them.
With write access: Add notes to incidents.
Analysis findings
Can read: Findings by area, so you can ask what to fix first.
How the WebPixie MCP server works
MCP (Model Context Protocol) is a standard way for an AI assistant to connect to other tools. WebPixie runs an MCP server at https://app.webpixie.io/mcp, so an assistant you already use can read your monitoring data and, if you allow it, change it.
You add the address to your assistant and sign in with WebPixie. A consent screen shows which assistant is asking, and you choose the workspaces it can reach (all of them or a selection) and read-only or read and write access for each. Each workspace can also be narrowed to specific sites and uptime monitors.
Once connected, the assistant works with your WebPixie data: uptime checks, SSL and domain expiry, DNS, links, sitemaps, incidents, and the findings from WebPixie’s site analysis. Because it can read all of that together, it can also help you decide what to act on: explain a finding, rank open issues, or suggest what to fix first. Access follows the same rules as the WebPixie API, so your workspace role and read-only settings apply.
For setup guides, permissions, and example prompts, see the MCP docs.
Ask in plain language
Uptime, SSL, domains, and incidents
Ask which sites have an SSL certificate that expires soon, why a monitor went down, or what WebPixie’s analysis found on a site. The assistant reads the data in your account, including the details behind an incidentand can help you work out what to do next. You get the answer in the chat instead of opening several dashboard pages.
Choose what the assistant can reach
Workspaces, access mode, and scope
Connect all of your workspaces or only some, and set each one to read-only or read and write. A read-only connection does not offer the tools that make changes. For finer control, narrow a workspace to specific sites and uptime monitors. When an assistant has write access, it is instructed to ask you before deleting anything.
Manage your monitoring from the chat
Sites and uptime monitors
With write access, ask the assistant to add a site, create or update an uptime monitor, or pause one during a deploy. The changes show up in WebPixie like any other. AI assistants can make mistakes, so check important changes before relying on them.
Works with the assistants you use
AI platforms and tools such as Claude, ChatGPT, Codex, and VS Code
Add https://app.webpixie.io/mcp as an MCP server or custom connector in your assistant, sign in with WebPixie, and approve the access you want. There are no API keys to create or copy.
Everything you need to monitor a website. In one workspace.
A quick look at other WebPixie features.
Why teams connect their assistant to WebPixie
No keys to manage
You sign in with your WebPixie account. There is no API key to create, copy, or rotate.
Answers come from your own data
The assistant reads your monitoring data and analysis findings, so its answers come from your account instead of general advice.
Available on every plan
MCP access is included on every plan, including Free, so you can try it on your own sites.
Frequently Asked Questions
Common questions about the WebPixie MCP server.
The WebPixie MCP server lets an AI assistant connect to your WebPixie account so you can ask about your monitoring data in plain language. MCP (Model Context Protocol) is a standard way for assistants to use other tools, and WebPixie runs its own MCP server. After you add the server address to your assistant and sign in, the assistant can read uptime data, SSL and domain expiry, DNS, links, sitemaps, and incidents from the workspaces you approve. If you allow it, it can also make changes such as pausing a monitor. The server uses the same data and rules as the API integration, so your workspace role applies, and it covers the same areas you see in uptime monitoring and incident management. For setup guides, see the MCP docs.
Claude (claude.ai and Claude Code), ChatGPT, Codex, and VS Code can connect to the WebPixie MCP server. Each one lets you add an MCP server or custom connector by address: add the WebPixie server address, sign in with your WebPixie account, and approve the access you want on the consent screen. If the platform or tool you use is not on WebPixie’s allowed client list, contact us and we will make it available for you quickly. Connecting does not require a paid plan, so you can try it on a Free workspace and then compare what other plans add on the pricing page. Once connected, the assistant works with the workspaces you selected and follows your role in workspace management. Step-by-step guides for each assistant are in the MCP docs.
No, the WebPixie MCP server uses OAuth sign-in instead of API keys. When you add the server address to your assistant, you are sent to WebPixie to sign in and approve access on a consent screen, and the assistant receives its access from that approval, so there is nothing to create, copy, or rotate. API keys belong to the API integration, where scripts and services authenticate with Basic auth. The two are separate: an API key does not work with the MCP server, and an MCP connection does not create an API key. Both follow the same access rules, including your workspace role and read-only settings, so choosing between them mostly comes down to whether you are talking to an assistant or running code. You can find the steps in Connect Your Assistant.
Only if you allow it. You choose read-only or read and write for each workspace on the consent screen. With read-only access the assistant cannot make changes and does not see the tools that would. With read and write access it can do things such as pausing or resuming a monitor, and it is instructed to ask you before deleting anything. You can also narrow a workspace to specific sites and uptime monitors, so an assistant only reaches the resources you pick. Account, billing, and team settings stay in the dashboard and are not available to assistants. Anything an assistant changes follows the same rules as the API integration, and the effects show up in your uptime monitoring like any other change. AI assistants can make mistakes, so review important changes before relying on them. For every choice on the consent screen, see Access and Permissions.
Every plan includes MCP access, including Free. You do not need a paid plan to connect an assistant. What the assistant can see and change depends on your workspace, so plan limits such as the number of sites and monitors, check intervals, and data retention still apply to the data it reads. If you also want to automate monitoring from your own code, the API is part of the paid plans. You can compare what each plan includes on the pricing page, and see how scripts and services connect on the API integration page. Workspace roles also decide who can approve write access for an assistant, which is covered in workspace management.
A screen in the WebPixie dashboard for viewing and removing assistant connections is coming soon. Each connection belongs to your user and covers the workspaces you approved on the consent screen, and removing a connection stops the assistant from using them on its next request. Until the screen is available, contact us if you need a connection removed. You can also change what a future connection can reach, because every new connection asks you to choose workspaces, access mode, and, if you want, specific sites and monitors. For how roles shape what an assistant can do, see workspace management. See Access and Permissions for how connections work.
Ready to ask about and manage your sites with your assistant?
Connect once, choose the access, and ask.