Last updated: 6 October 2026
Quick Answer: Zapier pricing starts at nineteen dollars ninety-nine a month for Professional with 750 tasks, billed annually. Team is sixty-nine dollars a month for 2,000 tasks and 25 users, Free covers 100 tasks, and Enterprise is quoted by sales. Growing teams then pay for tasks, AI steps and Agents activity.
Zapier charges per successful action, so a Zap that reads cheap at 500 runs can cost several times more at 5,000.
Most teams meet Zapier pricing twice: when they pick a plan, and a year later when a few workflows have become load-bearing. Three things decide the bill: how many steps count, which model tier your AI steps use, and what calls Zapier from outside a Zap.
How much does each Zapier plan cost in 2026?
Four plans. Free includes 100 tasks a month at no charge. Professional and Team are priced by task tier, starting at 750 and 2,000 tasks, and Enterprise is quoted by sales. Annual prepayment is a third cheaper than monthly.
Zapier pricing plans as listed on zapier.com on 6 October 2026.
| Free | Professional | Team | Enterprise | |
|---|---|---|---|---|
| Price a month | $0 | $19.99 at 750 tasks, $49 at 2K, $89 at 5K, $129 at 10K billed annually; $29.99, $73.50, $133.50 and $193.50 billed monthly | $69 at 2K tasks, $119 at 5K, $169 at 10K billed annually; $103.50, $178.50 and $253.50 billed monthly | Quoted by sales, per the plan selector |
| Tasks | 100 a month | from 750 a month | from 2,000 a month | annual task limits |
| Users | 1 | 1 | 25 | Unlimited |
| What you get | Two-step Zaps, Tables, Forms | Multi-step Zaps, premium apps, webhooks | Shared Zaps, SAML SSO | Admin controls, observability, account manager |
The task tier, not the plan, is what you are really buying, and both paid plans climb the same ladder to 2M tasks a month.
What do Zapier's Professional, Team and Enterprise plans include?
Professional buys capability, Team buys governance, Enterprise buys administration.
What does Zapier Pro pricing look like at higher task tiers?
Zapier Pro pricing climbs with the tier you select, not with the features, which stay the same all the way up. The ladder above roughly doubles between the 2K and 10K tiers while cost per task falls, so the cheapest way out of an overage is usually a higher tier.
What does Zapier Team add over Professional?
Team raises the ceiling to 25 users and adds shared Zaps and folders, shared app connections, SAML SSO and priority support. The real difference is ownership: on Professional, Zaps belong to whoever built them, so an engineer leaving takes a working automation.
What is in Zapier Enterprise pricing?
Zapier Enterprise pricing is quoted rather than listed. The plan card names unlimited users, advanced admin and app controls, annual task limits, observability and a technical account manager. The annual limit matters most: a seasonal workload can spike in one month without tripping an overage.
How do Zapier tasks work, and what counts as one?
A task is any successful action, and nothing else counts. That rule is more generous than teams assume, because the logic around an action is free. Zapier's task usage rates page sets the rate for every step type.
Counted, at 1 task each: every successful action step, every step inside a sub-Zap, the call and return steps around it, a code step on standard runtime, and a search set to proceed when nothing is found.
Free, at 0 tasks: all triggers, including polling triggers; Filter and Paths; Formatter, Delay, Looping, Digest, Storage and Zapier Manager; Tables and Forms steps; and any step that errors, halts or never runs.
How much do AI and MCP steps cost in tasks?
AI by Zapier is metered by model tier, and the multiplier applies twice: to the step itself and to every tool call the model makes. Standard costs 1x, Advanced 3x and Premium 5x, and connecting your own model provider account brings it back to 1x. New AI steps on a paid plan default to Premium, so a step nobody configured is the most expensive one you own.
Zapier MCP pricing draws on the same pool rather than a separate fee: each successful tool call from an outside AI client costs 2 tasks, and failed calls are free.
What do Zapier Agents and Chatbots cost on top of a plan?
Both are add-ons with their own meters. Zapier Agents pricing is measured in activities, not tasks: the free tier covers 400 activities a month, Agents Pro costs $400 a year, which is $33.33 a month, for 1,500 activities, and Enterprise is custom; Zapier chatbot pricing is a separate add-on again, listed on the same plan selector at $160 a year for Chatbots Pro and $800 a year for Chatbots Advanced.
An activity is any billable thing an agent does: firing a trigger, answering from a knowledge source, running an action, browsing a page, running a web search, or taking a Chrome extension message. Activities and tasks do not draw on each other, so a team running both manages two budgets.
What happens to your Zapier bill when Zaps run at volume?
You switch to pay-per-task automatically. On paid plans, exceeding the task limit does not stop your Zaps: Zapier bills the extra tasks for the rest of the cycle, at a rate its billing documentation says depends on your plan and billing cycle type and is shown in Billing settings. Pay-per-task stops at 3 times the plan's limit, and new runs are then held until the cycle resets.
Work an example. Take a six-action Zap where three steps are app actions and the rest are a Filter, a Paths branch and a Formatter. Only the app actions count, so each run costs 3 tasks. Run it 3,000 times a month and you use 9,000 tasks, which clears the 5K tier and lands on the 10K tier priced above.
When is a custom AI workflow cheaper than Zapier?
When the multiplier, not the plan, is what you are paying for. Four signals say the crossover is close: AI steps on the 5x Premium tier at volume, many counted actions per run, agents acting on internal systems that need audit trails, and data you cannot send to a shared cloud. Each scales by run count; a custom build scales by compute.
Model inference is usually the cheap part. On OpenAI's API pricing, read on 6 October 2026, gpt-6-luna costs $0.10 per million input tokens and $0.50 per million output. At 9,000 runs a month using 2,000 input tokens each, that is 18 million tokens, a rounding error next to a 10K task tier.
Counting units, not headline prices, are the honest comparison across tools. n8n meters cloud plans by workflow execution, so a long workflow is one unit; Make counts one credit per module action, closer to Zapier. Our guide to n8n alternatives covers that, and the build-or-buy question is in custom AI agents versus no-code builders.
What mistakes should you avoid before upgrading a Zapier plan?
Upgrading on instinct is expensive. Five checks before you move a tier:
- Read Task History by Zap, not the account total. One Zap usually owns most of it.
- Open every AI by Zapier step and check its model tier. The Premium default costs five times a Standard one.
- Confirm which steps are actually counted. Teams buy tiers they do not need after assuming filters were billable.
- Count users before features. If you need shared ownership rather than throughput, Team at a low tier beats Professional at a high one.
- Decide pay-per-task deliberately: on for critical Zaps, off where a held run beats a surprise invoice.
What should a quote for a custom AI workflow include instead?
Ask for the same line items every time: the processes in scope and their volumes, the systems being integrated, where data is processed, the run cost at target volume, who operates it after handover, and what rollback looks like. Our breakdown of AI workflow cost sets out what each line contains.
How Origins AI replaces high-volume Zaps with custom AI workflows
Origins AI (originshq.com) is a US-based AI engineering and workflow development partner for product teams. The work here usually starts with one Zap rather than a rebuild: the workflow that has outgrown the meter moves onto owned infrastructure, and the rest stays put.
The agentic automation page describes the shape: approval workflows, triage and routing, monitoring and response, and data operations, rolled out through process mapping, agent design, a pilot on one process, then scale and monitor. Origins AI reports an average of 1,200 hours saved a year and decision cycles moving from 45 minutes to 3 minutes. Those are company figures.
Origins AI does not publish a rate card, so there is no like-for-like number to set against a task tier. The AI workflow development page lists the engagement side, scoped per project.
Talk to an engineer
If one workflow owns most of your tasks, that is the one worth costing properly. Book a call and we will work through the numbers.