Zapier pricing, and the open source alternative you can own

How Zapier's task-based pricing works

Zapier bills by the task, and a task is one successful action an automation completes. A Zap's trigger and a check for new data cost nothing, so a workflow pays only when it runs an action. The same task pool now covers Zap workflows, AI steps, code, MCP, and the SDK. Zapier's plans start with a free tier at $0 for 100 tasks per month, then Professional from $19.99 per month and Team from $69 per month for 25 seats, and a higher task tier lowers the cost per task as volume grows. Enterprise is quoted on request. Every figure on this page was checked October 2026.

What n8n and Make change about the bill

n8n bills an execution, which is one full run of a workflow no matter how many steps it holds, and its pricing starts at Starter for $20 per month billed annually, Pro at $50, and Business at $800 on a self-hosted instance. Its Community Edition is a self-hosted version you run yourself at no cost. Make bills credits instead, where most module actions use one credit; its plans run from a free tier of 1,000 credits per month to Core at $9, Pro at $16, and Teams at $29. Make is hosted on its own cloud, so there is no self-hosted floor under its bill.

When a hosted tool is the cheaper call

A hosted tool is the right answer when volume is low and the workflow should not become your team's problem. The free tiers at Zapier and Make, and n8n's self-hosted Community Edition, cover a few simple automations, and the monthly cost stays small while the count of runs stays small. The shape to watch is that the bill tracks usage. Zapier bills each successful action and Make bills each module action, so a workflow that fans out across many steps or many records grows the bill with it. The plan that was cheap at fifty runs a month is not the plan you are on at fifty thousand.

When self-hosting changes the cost curve

Self-hosting an open-source system trades a subscription for infrastructure and ownership. You pay for the machine and the upkeep, and in return the software layer stops scaling with your usage. n8n lets you run its Community Edition yourself, and Kortix goes further: self-hosting the whole system is free, and the configuration is a git repo you own. Self-hosting starts with two commands:

curl -fsSL https://kortix.com/install | bashkortix self-host start

Ownership matters once automations stop being a side tool and start carrying real work, because cost then follows the compute you use rather than the number of steps a workflow happens to contain. The open-source Zapier alternative page covers what that setup takes.

Why Kortix is the open source option to price against

Kortix prices two ways, usage and seat, and the system is open source. The free tier is $0 with 200 credits each month for sandbox compute and one project. The team plan is $40 per seat per month with 2,500 credits per month per seat, pooled across the team, and self-hosting the system is free. Model access is your choice: bring your own API key or connect a ChatGPT subscription you already pay for, so that part of the bill stays yours. Agent Computer compute is billed per second and per resource, about $0.20 per hour for the default machine, and $0 while a session is stopped. The Zapier alternative overview collects the wider field if you are still choosing.