Claude MDN connectorWhat Claude can do in your MDN account.
The Claude MDN connector exposes 3 tools, all 3 of them read tools. 0 write, 0 undocumented. Claude searches Mozilla's web docs, pulls a full page and checks browser support. Below: what you get out of it, and the caveats Mozilla attaches.
Verified Trustpilot reviews · AI, automation & growth agency
Current docs instead of a stale memory
A language model learned the web up to some date, and the web kept moving. With MDN connected, Claude looks up the reference page at the moment you ask: a CSS property, a JavaScript method, a browser's support status. The answer rests on what MDN publishes today, not on what the model happened to absorb.
Settle the "does this work in Safari?" debate. get-compat retrieves the browser compatibility data MDN publishes, useful before adopting a recent CSS feature.
Land on the right page without digging. Describe what you need in plain words ("cancel a fetch request"), search scans MDN, and Claude points you to the relevant articles.
Read the whole reference. Once the page is found, get-doc brings it back as markdown, right in the chat.
Mozilla says it aims to give language models and coding agents a reliable, up-to-date source on the web platform.
What the connector won't do: it touches no account. By default its sheet only exposes reading, and nothing fires on its own when MDN ships a new page. Mozilla also runs it as an experiment that can be withdrawn at any time. If you came here looking for scheduled, hands-off workflows, the Integrations hub covers automation tools, which are a different job.
Five terms before you connect MDN
Defined once, used as-is afterwards.
- Connector
- The link you set up once between Claude and a service, so Claude can use it while it answers you.
- Tool
- One named capability a connector opens to Claude. Claude picks the ones it needs on its own; the directory lists them by name.
- Authorization
- The service's own screen where you grant Claude the access it will use. Given once per person, revoked the same way.
- Approval
- The confirmation Claude waits for before an action that changes an account, shown in the chat right when it matters.
- MCP
- The shared standard behind connectors: it's what lets an assistant like Claude talk to an outside service.
Connect the Claude MDN connector
- 01
Find MDN in your connectors
In Claude, open Customize, then Connectors, and look for MDN. On a Team or Enterprise workspace, an Owner or Primary Owner enables it for the organization first, then each member connects individually.
- 02
Start the connection
Click Connect on the MDN row and follow the window the service opens itself. If the link ever breaks, use Disconnect, then connect it again from the same row.
- 03
Review the authorization screen
Read the authorization screen before accepting. It belongs to the service, and it decides what the access covers, not Claude. You can also withdraw that access later from the service side.
The 3 tools in the Claude MDN connector
MDN gives Claude 3 tools: 3 that read your account, 0 that change something in it.
Three lookup tools, described by the tool list Mozilla's service publishes. Names stay in English, exactly as Claude shows them.
- 3 read
What Claude reads (3)
3 toolsThree tools that look things up in MDN's public documentation.
get-compat
Pulls MDN's browser compatibility data so you know where a web feature is supported. Mozilla presents the service as an up-to-date source of web platform documentation.
get-doc
Fetches one MDN documentation page as markdown, straight into the chat. It comes in once the right page is found, for instance with search.
search
Runs a query across MDN for a web technology and returns the documents that match your question. Claude uses it to locate the right page before reading further.
What Claude asks first, and who can switch it off
By default, Claude stops and asks for your confirmation before each action it takes on an account for you. Here the three tools only consult MDN, and no source describes an approval specific to any of them.
On Team and Enterprise, workspace owners decide whether members may let certain actions through without being asked every time. They can also restrict a connector for the whole organization. For instance, keep reading open and close every write, with a setting that applies to every member and that nobody overrides from their own account. Claude also works with the connected person's rights and nothing beyond them: what you can't reach yourself, Claude can't reach either.
Which plans include it
None of the 819 directory sheets shows plan availability. That information isn't published anywhere on a per-connector basis.
The general rule is public: remote connectors are open to all users on Claude, Cowork, Claude Desktop and mobile. On Team and Enterprise, an Owner or Primary Owner switches the connector on first. For the current status on your own account, the only reference is MDN's sheet in the official directory, which the button on this page opens. Any other page quoting plans for this connector cites no source for it.
Where MDN stops inside Claude
A connector isn't an automation: Claude calls these tools while answering you, and nothing runs in the background. Only Mozilla, the publisher, documents this connector.
The sheet exposes reading only, which is an observed floor, not a promise: an admin can open more than any public page shows. The directory badge isn't a security audit, and Anthropic states on every sheet that it doesn't control a publisher's tools. A connected community connector has exactly the same access as any other. So only plug in what comes from a publisher you know, Mozilla in this case. Other connectors are gathered on the Claude connectors page.
Need help connecting MDN to Claude?
A person reads every message.
