Connectors & MCP
This is where Claude stops being a separate window and becomes part of your workflow — reaching into your mail, files and calendar. Connect the safe things first, review every permission.
What connectors unlock
One-click integrations (built on MCP) that let Claude read from and act in your work tools. Available on all plans (Free is limited to one custom connector); they work across chat, desktop, Cowork and Code. A few things they make possible:
"Summarise the supplier email thread and draft a reply."
"Find last month's report in Drive and pull the headline numbers."
"Pull this query → build a Canva deck → create an Asana task," in one conversation.
Interactive: the Connector Map
Dozens of reviewed connectors exist. Tap one to see what it does — and, crucially, the permission you'd be granting.
Select a connector to see what it enables and what you're granting.
MCP — "a USB-C port for AI"
The Model Context Protocol is Anthropic's open standard for connecting Claude to external tools and data. The plain-English point for a business audience: it's why Claude connects to so many systems, and why the list keeps growing. You don't implement it — you benefit from it, the same way you don't need to understand TCP/IP to send an email.
The privacy & governance gate
Connecting grants access — each connector has read-only vs. write/delete permissions you review at authorisation. Use a personal or sandbox account for the exercise, not a locked corporate one, unless IT approved it. Custom (unverified) connectors carry prompt-injection risk. On Team/Enterprise, admins gate which connectors are available — that's the right path for work data.
Key takeaway
Connectors put Claude inside your stack; MCP is why that stack keeps growing. Connect the safe things first, review every permission, keep work data on the governed plan.