AI-discovered actions
The agent searches Zapier’s catalog of more than 9,000 apps and 66,000 triggers and actions to find a suitable tool for a natural-language request.
Zapier MCP
Freemium
Zapier MCP connects AI clients such as Claude, ChatGPT, and Cursor to actions across Zapier’s app ecosystem. It lets an agent discover an appropriate action, request missing details, and execute it with authentication and permissions managed through Zapier.
What is Zapier MCP?
Zapier MCP is an AI action layer that connects clients such as Claude, ChatGPT, and Cursor to business apps through Zapier. An agent can interpret a request, search more than 9,000 apps and 66,000 triggers and actions for a suitable tool, ask for missing details, and execute the action with authentication and permissions managed by Zapier.
Users set it up by adding Zapier MCP to a supported AI client and selecting the apps it may use. Existing Zapier app connections appear automatically, and new connections can be added. A completed interaction can remain a one-off action, be saved as a reusable Skill, or be deployed as a Zap for independent execution.
What can Zapier MCP do?
Client connections
Zapier MCP can be added to AI clients including Claude, ChatGPT, and Cursor, with the page also referring to other supported AI clients.
Managed app access
Users choose which apps the connection can use. Existing Zapier connections appear automatically, and new app connections can be added.
Missing-detail prompts
When a request does not contain enough information, the agent can ask for the missing details before running the workflow.
Code for nonstandard tasks
When a standard action is not a fit, Zapier says MCP can write and run code to reach an internal API or reshape data.
Reusable execution
Useful interactions can be saved as Skills or deployed as Zaps that run independently, turning an interactive action into a repeatable workflow.
Use Cases
“Act on requests in an AI chat”
An individual can describe a desired business action in Claude, ChatGPT, or another connected client instead of manually selecting an integration step for every request.
“Work across connected business apps”
A team can allow an agent to use selected apps such as Gmail, Slack, or Salesforce, then have it discover the relevant action from the available connections.
“Handle internal or irregular data work”
When the required operation does not match a standard app action, MCP can use code to reach an internal API or transform data during the request.
“Turn successful prompts into automation”
A workflow that proves useful interactively can be saved as a Skill for reuse or deployed as a Zap when it should run independently.
“Manage AI access at organizational scale”
Organizations can pair MCP with Zapier’s broader team and enterprise controls, including shared app connections and administrative access controls listed on the pricing page.
Frequently Asked Questions
Which AI clients can connect to Zapier MCP?
The product page explicitly names Claude, ChatGPT, and Cursor, and says MCP can connect to other supported AI clients. It does not provide a complete client list on the supplied page.
How are apps selected for an MCP connection?
Users choose the apps the connection can use. Apps already connected to their Zapier account appear automatically, and they can connect additional apps.
Does the agent choose the action automatically?
Zapier says the agent searches its catalog of more than 9,000 apps and 66,000 triggers and actions to find the right tools for the stated goal. It can also ask for missing details before executing the workflow.
Can a one-off MCP action become a recurring automation?
Yes. Zapier says useful workflows can be saved as reusable Skills or deployed as Zaps that run independently.
How is MCP usage priced?
Zapier’s pricing page describes task-based pricing across its platform. MCP, Zaps, AI steps, code, and SDK usage draw from a shared task allocation, and usage can vary with the connector type and workflow complexity. The pricing page lists a free tier and paid plans.
Quick Facts
- Product type
- AI-to-app action and automation layer
- Named AI clients
- Claude, ChatGPT, and Cursor
- App coverage
- More than 9,000 apps
- Available triggers and actions
- More than 66,000
- Workflow outputs
- One-off actions, reusable Skills, or independently running Zaps
- Pricing model
- Task-based usage with a shared allocation across Zapier products
Zapier MCP Alternatives
Kody
kody.codes
Kody is a software platform that gives connected AI agents shared memory, credentials, reusable packages, integrations, and scheduled execution. It helps developers carry work between agents such as Cursor, Claude, ChatGPT, and Codex without rebuilding the same automation each time.
Any AI for Notion
anyaifornotion.com
Any AI for Notion connects ChatGPT, Claude, Gemini, and other AI agents to a Notion workspace. It lets users search, read, create, and update Notion pages and databases from mobile or desktop using their own AI credentials.
Edgee
www.edgee.ai
Edgee is an agent gateway for coding teams that reduces token usage, routes requests across models, and provides usage visibility. Its Compression V2 works as a drop-in CLI layer for Claude Code, Codex, OpenCode, Cursor, and other supported agents.
Flocker
flocker.md
Flocker is an AI agent management platform for creating agent profiles, coordinating multi-agent work, and monitoring activity from a shared dashboard. It is designed for people running multiple agents across tools such as Claude Code, Codex, OpenClaw, OpenCode, and MCP-compatible workflows.
Pancake
getpancake.ai
Pancake is an AI agent platform for founders and small teams, handling operational work across the tools they already use through always-on, multi-channel access.
Automat Workforce
runautomat.com
Automat Workforce provides Ace, an AI teammate with its own Mac, email, phone, and company account. It works through workplace tools such as Slack, Teams, Gmail, Microsoft 365, calendars, chat, phone, SMS, and iMessage to complete tasks and request human input when needed.