Claude integrations: How to use Zapier with Claude (Fable 5.2, Opus 5.5, and more)

With every new model release, Claude tops almost every benchmark. And chatting with Claude makes it clear why: it's good at everything.

But while there's power in a back-and-forth conversation, you can accomplish a lot more when you use Zapier to connect Claude to the rest of your apps and let automation carry out entire workflows for you.

Here's how it works.

Table of contents:

• Build and automate workflows from Claude

• Add Claude's brain to your existing workflows

• Claude automation FAQ

Build and automate workflows from Claude with Zapier MCP

Install Zapier MCP in Claude and it can reach 9,000+ apps. Connect it once, tell Claude which apps and actions you want it to use, and it works across your whole stack, straight from the chat window. Zapier manages the logins, so Claude never handles your API keys.

Here are a few things you might tell Claude to do once Zapier MCP is connected:

• "Pull my last five Salesforce opportunities, calculate weighted pipeline, and push a summary to Google Sheets."

• "Catch me up on the Slack messages I missed this morning, flag anything that needs my attention, and add those items to my Asana task list."

• "Draft a follow-up email based on the last call note in HubSpot and create a calendar invite for next week."

None of those require you to open a single other app—it all happens right in Claude.

Claude's built-in connectors cover the basics, like looking up a record or updating a field. Zapier goes further, so Claude can do things like convert a lead in Salesforce or enroll a contact in a HubSpot nurture sequence. And because your connections live in Zapier, they work from ChatGPT, Cursor, or any other AI tool that supports MCP. You don't start over when you switch.

Turn what you repeat into a Zap

One-off requests are just the start. When you find yourself asking Claude for the same thing again, ask it to build a Zap instead: "Every time a new lead fills out our HubSpot form, enrich it, score it, and post the hot ones to #sales." Claude drafts the Zap and deploys it to Zapier, where it starts on real events, like a new lead, a ticket, or an email reply, instead of waiting for you to ask.

A few things make this hold up over time:

• Same result every time. Zaps use AI only for the steps that need judgment. Everything else is a regular step, which runs faster and costs less.

• Fixes what breaks. Every run is logged. When one fails, Zapier finds the problem and proposes a fix, which you can approve or work through with Claude.

• Cheaper and sturdier over time. When an AI step does the same thing every time, Zapier can offer to make it a regular step: same result, fewer tokens.

Add Claude's brain to your existing workflows with AI by Zapier

AI by Zapier lets you add a Claude step wherever you need it, right inside your workflow. You don't even need a separate Anthropic account to use it.

This means Claude only runs in the steps that need judgment, and regular steps handle everything else, so you aren't burning through tokens on tasks that run the same way every time. For example, if you're sending data from a form to a spreadsheet, that's a job for deterministic automation. But if you need to summarize a customer complaint, classify a support ticket, or draft a personalized email based on CRM data, that's when you get Claude involved.

Here's what that looks like in practice: say a new row gets added to a Google Sheet whenever someone submits a feedback form. Instead of reading every response manually, you set up a Zap that starts on new rows, sends the text to Claude with a prompt like "Summarize this feedback and classify it as positive, neutral, or negative," and then posts the result to a Slack channel. The whole thing runs in the background, and Claude only fires when there's actually a row to process. Because the Zap starts when something happens, it keeps working even when your laptop is closed.

You also get complete control over the Claude step: you can add knowledge sources, include agentic tool calling, and continue testing until it looks just right.

You tell Claude or Zapier which model you want to use, and the cost comes out of your Zapier task count, which means no tokens and no separate Anthropic bill. And if you decide the Claude model isn't working the way you want, you can swap it for any other model you want, without rebuilding the workflow. (You also have the option to bring your own key [BYOK] if you want to use a model that's not available in AI by Zapier—in that case, you'll pay your provider directly.)

Here's a quick look at the Claude models currently available on Zapier; every time a new Claude model is released, it gets added to the lineup immediately.

If you're not sure which model to reach for, check out AutomationBench, Zapier's benchmark for evaluating AI models on real business workflows. It's the quickest way to see how Claude's models compare on the tasks you'll actually be running.

Claude automation use cases

Here's a look at how you might use AI by Zapier to bring Claude into your workflows:

• Write blog posts and outlines based on keywords. Trigger a Zap when a new keyword lands in a Google Sheet (or wherever your team tracks ideas), send it to Claude with a detailed system prompt and access to your style guide, and get a full draft or structured outline delivered to Notion, Google Docs, or your CMS of choice.

• Summarize key information. Schedule a Zap that summarizes the previous day's Gong calls every morning and posts highlights to Slack.

• Turn content into social posts. Take a new blog post, product update, or press release, and have Claude draft platform-specific copy for LinkedIn, X, and Instagram in a single step, then send it directly to your social media management platform.

• Analyze surveys and feedback. When responses come in from Typeform, Google Forms, or another survey tool, pipe them to Claude for sentiment analysis, theme extraction, or NPS categorization—then push the results to your database to build a running feedback dashboard.

• Escalate issues automatically. Have Claude read incoming support tickets or incident reports, assess urgency against a defined rubric, and trigger a Slack alert or calendar invite for anything that crosses your severity threshold—without anyone triaging every ticket manually.

• Route creative assets for review. When a new asset lands in Figma, Dropbox, or Google Drive, Claude can evaluate it against a checklist, add a summary comment, and send it to the right reviewer based on asset type—compressing the review cycle for high-volume creative teams.

Claude automation FAQ

Why use Zapier if I already have Claude?

Keep Claude. Zapier makes it do more. Claude's built-in connectors cover a handful of popular apps. Zapier connects it to your whole stack and keeps your Zaps running after the chat ends. If a Zap breaks, Zapier finds the problem and proposes the fix.

Can't Claude's scheduled tasks do this?

They can, but it's the long and expensive way. A scheduled task runs only at a set time, and the model works through every step each time. A Zap starts when something happens, like a new lead or a new ticket, and uses AI only in the steps that need it. It runs exactly when it's triggered, gives you a reliable result, and costs less.

Why not have Claude Code write me a script?

You can, and the first version will probably work. The trouble starts after. A script needs somewhere to run, a way to hear about events, and someone to keep it alive. When an API changes or a login expires, it quietly stops, and you have to fix it. It also holds your keys, usually on a laptop or in a repo. With Zapier, Claude builds the Zap, Zapier hosts it, runs it on real triggers, and keeps your connections working. When something breaks, Zapier finds the problem and proposes the fix, and you approve it.

How does Zapier MCP work with Anthropic's Claude?

Zapier MCP connects Claude to 9,000+ apps. Once you set it up, Claude can call actions—like reading a Salesforce record, posting to Slack, or creating a Google Calendar event—directly from the chat window. You define which apps and actions Claude has access to when you configure the MCP server in Zapier, so you stay in control of what it can do. Claude treats each authorized action as a tool it can invoke mid-conversation, meaning you can give a single prompt like "Pull my last five opportunities from Salesforce and send a summary to the team Slack channel," and Claude handles the whole sequence without you switching apps. If it's work you'll repeat, Claude can turn it into a Zap.

How do I connect Zapier to Claude via MCP?

Go to zapier.com/mcp to get started. Zapier generates a unique MCP server URL that you paste into Claude's settings under Connectors. Once connected, Claude will see your authorized Zapier actions as available tools. Select the specific actions you want to expose on Zapier—you don't have to grant access to everything—and Claude can start using them right away from any conversation.

Automate Claude with Zapier

Whether you want Claude inside your workflows or building and running them from chat, Zapier is where the work keeps running. Either way, Zapier manages the connections, so Claude gets access to your apps, not your keys.

Related reading:

• AI by Zapier: Easily add AI steps to your workflows

• How to use the Claude API

• Claude connectors: How to connect Claude to other apps

This article was originally published in November 2023 by Elena Alston. It was most recently updated in November 2026.