Plug Claude into Gmail, Drive, Slack, and the rest
It is late Friday afternoon. You need release notes for this week's sprint by end of day. Normally you would open Linear, read every completed ticket, find the template on Google Drive somewhere, copy-paste, rewrite, hope you did not miss anything. Instead you type one sentence into Claude: "Pull the completed tickets from this week's production sprint and create release notes using our standard template." A minute later, they are done — with real ticket titles, real descriptions, the right formatting. This is what connectors unlock.
Connectors are integrations that give Claude access to the tools and data you already use. They turn Claude from a helpful assistant into an informed collaborator working with your actual information, not a guess at it.
The universal standard underneath: MCP
Before you wire anything up, it helps to know what is happening under the hood. Every connector speaks a common language called the Model Context Protocol, usually shortened to MCP.
The cleanest analogy is hardware: MCP is USB-C for AI. Before USB-C, every device needed its own cable and its own port. Plugging a new device in meant buying the right adapter, installing the right driver, and hoping they all played nicely. USB-C collapsed that into one standard — one cable, works with everything.
MCP does the same for AI. Instead of Claude needing bespoke code for every tool on the planet, any developer can build a connector that speaks MCP, and it plugs straight into Claude. That is why the connectors directory can keep growing without each new tool requiring a custom integration.
Two kinds of connector
You will see two categories in the directory.
Web connectors link Claude to cloud services — Gmail, Google Drive, Notion, Slack, Asana, Linear, Stripe, Salesforce, and many more. They work from claude.ai in your browser, and they cover most of the tools knowledge workers live in.
Desktop extensions run locally on your computer through the Claude Desktop app. They give Claude access to local files, native applications, and things only available on your machine — reading and organising documents, controlling your browser, integrating with design tools like Figma.
The short rule: web connectors for cloud tools, desktop extensions for local tools.
Step 1: Browse the directory
Anthropic maintains a directory of recommended connectors at claude.ai/directory. It is organised into two tabs — Web and Desktop extensions — and grouped by category: productivity, communication, development, business, automation.
You can also get to the directory from inside any chat. Click Search and tools in the lower-left of the chat window, then Add connectors. Same directory, different entry point.
Step 2: Set up your first web connector
Pick one tool you use every day. Here is the flow.
- Find the connector. Open claude.ai/directory or click Search and tools → Add connectors in any chat.
- Click Connect. Select the connector you want.
- Authenticate. Claude redirects you to the service's login page. Sign in with your existing credentials.
- Grant permissions. Review the specific permissions Claude is requesting, then authorise access. Claude explains what each permission unlocks before you approve it.
- Test the connection. Go back to Claude and try a simple request — "Can you access my [tool name]?" is the quickest way to verify it is live.
Once connected, Claude can search, read, and in some cases take actions inside that service — depending on the permissions you granted.
Step 3: Set up a desktop extension (if you need one)
Desktop extensions live inside the Claude Desktop app, not the browser. To install one:
- Download and install the Claude Desktop app if you have not already.
- Open Settings → Extensions inside the app.
- Browse available extensions and click Install on the one you want.
- Follow the extension-specific setup. Some want a folder, some want a tool, some want credentials.
Desktop extensions include things like local file access, browser control for automated web tasks, and native application integration. They are the right choice when the work lives on your computer, not in the cloud.
Step 4: Use your connected tools in practice
Once tools are connected, Claude considers them automatically when relevant. You do not need to say "use the Google Drive connector." You just ask. A few patterns that land well:
Project management (Asana, Linear, Jira).
- "What are my highest priority tasks due this week?"
- "Create a new task for reviewing the Q4 budget proposal."
- "Summarise the status of our product launch project."
Communication (Slack, Gmail).
- "Find the email thread where we discussed the vendor contract."
- "Draft a reply to the latest message in the #marketing channel."
- "What did the team decide about the timeline in yesterday's discussion?"
Documentation (Notion, Google Drive, Confluence).
- "Search our documentation for our brand voice guidelines."
- "Summarise the meeting notes from last week's product review."
- "What does our style guide say about using contractions?"
Business tools (Stripe, Salesforce).
- "Show me revenue trends for the past quarter."
- "What's the status of the Acme Corp opportunity?"
- "List recent transactions over $1,000."
When you want to be explicit, tell Claude which tool to use — "pull this from Google Drive" or "check the #launch channel in Slack". Otherwise let Claude decide which connector to reach for based on the question.
Security and permissions — the things to know
Granting Claude access to an external service is not trivial. Four principles keep you safe:
- Scoped access. Permissions are specific to what the connector needs. You can toggle individual permissions on and off inside each application's menu.
- Claude sees what you see. A connector cannot reach data you do not have access to. Connecting your work email does not give Claude access to your CEO's inbox — only your own.
- Revocable at any time. You can disconnect a service through Claude's settings or through the third-party service's security settings. Nothing is permanent.
- Only install connectors you trust. Custom connectors (not in the Anthropic directory) can be built by anyone. Treat them the way you would treat any other installable software — review the source before you plug it in.
The bigger picture
Connectors are the moment Claude stops being a separate tab and becomes part of your workflow. Meeting notes become project tasks. Project updates become executive summaries. Scattered documents become one cross-referenced answer. The goal is not to connect everything — it is to connect the tools where your work actually lives, so Claude stops asking you to paste context it should already have.
Before you move on
Open claude.ai/directory right now and pick one connector for a tool you open every single morning. Connect it. Ask Claude one real question that uses it. That is the whole unlock.
Key Takeaways
- 1Connectors are integrations that give Claude access to your existing tools — they turn Claude from a general assistant into an informed collaborator working with your real data.
- 2The Model Context Protocol (MCP) is the universal standard behind every connector — think of it as USB-C for AI, where one standard lets Claude plug into many tools.
- 3There are two kinds: web connectors (for cloud services like Gmail, Notion, Slack, Linear) and desktop extensions (for local files and native apps via the Claude Desktop app).
- 4Setup is a short flow — browse the directory at claude.ai/directory, authenticate with your existing credentials, grant scoped permissions, and test with a simple verification question.
- 5Claude only sees what you can see, permissions are scoped and revocable, and you should only install connectors from sources you trust — treat them like any other piece of installable software.