Contents
- Verdict
- What is Zapier?
- Best for, and not for
- Strengths and limits
- Pricing
- What people actually build with Zapier
- What Zapier actually costs at small-business volume
- The pricing gotcha: held runs and the AI multiplier
- What breaks in production, and how you find out
- What it connects to that you already run
- Zapier alternatives: what to pick instead, and when
- Migration and lock-in: can you get your work out?
- FAQ
Zapier is the fastest way to get a simple automation live and the most expensive way to run a complicated one. Its published pricing as of August 2026 starts at $29.99 a month for Professional with 750 tasks included, scaling on a task slider up to $5,099 a month at two million tasks; Team starts at $103.50 a month with 2,000 tasks and 25 seats. Because it bills each successful action as a task, a six-step automation running 2,000 times a month is roughly 10,000 billable tasks, and the same work on a per-execution competitor sits on a plan costing a fraction of that. Buy Zapier for its connector catalogue and its ease, go in knowing the meter, and revisit it the moment volume climbs.
What is Zapier?
Zapier is one of the most established no-code automation tools, built around simple trigger-and-action workflows called Zaps. Its main strength is breadth: it connects to thousands of apps, so most business tools you use are likely supported. It has added multi-step Zaps, paths, and AI features on top of the classic two-step model.
Best for
Teams that want the widest app coverage for straightforward automations.
Not for
Any workload where volume multiplied by steps runs into five figures a month. At that point you are paying a premium for a connector catalogue you have already finished using, and a per-execution tool does the identical work for a fraction of the price.
Strengths
- The largest connector catalogue in the category by a wide margin, so the obscure app you depend on is probably supported
- Genuinely usable by a non-technical owner on day one, which is worth more than it sounds when nobody has time to learn a canvas
- Filters and Formatter steps do not consume tasks, so cheap pre-filtering is a legitimate cost strategy
- Failed actions are not billed, so a broken Zap costs you reliability but not money
- Mature error reporting, Zap history and replay, which is better than most competitors at telling you what went wrong
- Deep coverage of the small-business stack: Google Workspace, Microsoft 365, QuickBooks, Xero, Shopify, Stripe, Square, HubSpot, Mailchimp, Calendly, Slack and most CRMs
- AI actions, Zapier Agents and a widely used MCP endpoint, so it plugs into agent tooling without custom work
Limitations
- Per-task billing is the most expensive meter in the category once workflows have more than two or three actions
- Professional includes exactly one seat, so the moment a second person needs to edit Zaps you are looking at Team from $103.50 a month
- Multi-step logic, paths and loops feel bolted onto a two-step product and get awkward past moderate complexity
- Polling triggers on cheaper plans check infrequently, so 'instant' often is not, and the faster intervals sit higher up the ladder
- Premium apps are gated behind paid tiers, which can force an upgrade for reasons unrelated to volume
- No self-hosting, no data residency control, and no way to keep records inside your own network
Zapier pricing
Zapier bills per task, where a task is each successful action a Zap completes; triggers, polling, and built-in steps like Filter and Formatter do not count, and some AI actions consume multiple tasks. Published pricing as of August 2026: Free at $0 with 100 tasks a month and one seat; Professional from $29.99 a month billed monthly or $19.99 billed annually at the 750-task tier, on a slider that runs up to $5,099 a month for two million tasks, with one seat; Team from $103.50 a month or $69 annually at 2,000 tasks with 25 seats; Enterprise custom with unlimited seats. Annual billing is advertised at 33 percent off the monthly rate. There are no separate per-seat charges - seat counts are fixed by plan.
| Plan | Price | What you get |
|---|---|---|
| Free | $0 | 100 tasks/mo, 1 seat, two-step Zaps (as of Aug 2026) |
| Professional | from $29.99/mo | From 750 tasks/mo on a slider up to 2M tasks at $5,099/mo; 1 seat; $19.99/mo equivalent billed annually |
| Team | from $103.50/mo | From 2,000 tasks/mo, 25 seats; $69/mo equivalent billed annually |
| Enterprise | Custom | Custom task volume, unlimited seats, advanced admin, SSO |
Pricing reflects public plans as of August 6, 2026 and can change. Check Zapier for the latest.
What people actually build with Zapier
Zapier's sweet spot is narrow and genuinely valuable: a short automation between two apps, live the same afternoon, built by someone with no technical background. Here is what that looks like when it works.
Inbox triage for a small team. A Gmail trigger fires on new mail to a shared address, a Filter by Zapier step drops anything from a known newsletter sender for free, an AI step classifies the rest, and the result creates a HubSpot ticket. Filters and Formatter steps consume no tasks, so the discarded mail costs nothing - which makes filtering early the single most effective cost control on the platform.
Lead capture from Facebook Lead Ads. A new lead ad submission creates a contact in the CRM, sends an SMS through Twilio inside sixty seconds, and posts to a Slack channel so somebody can call. This is two or three billable actions and it is the case Zapier is best in the world at.
Booking-to-record for a dental practice. A Calendly booking creates a row in a spreadsheet, sends a confirmation email, and adds a reminder task. The value here is entirely that both apps are supported and nobody had to think about it.
Invoice chasing for a bookkeeping firm. A scheduled Zap reads overdue invoices from QuickBooks and emails each client a reminder. Note that the schedule trigger itself costs nothing, and only the emails bill - which makes the cost of this one predictable from the invoice count rather than the run count.
Connecting a niche system nothing else supports. The practice-management product, the trade-specific job costing tool, the regional booking platform. This is the actual reason to pay Zapier's premium and it is a good one; you are buying the long tail of the catalogue, not the builder.
What Zapier is genuinely bad at: anything with real logic in it, and anything at volume. Paths and loops feel bolted onto a product designed for two steps, and per-task billing means cost scales with both frequency and step count at the same time - the worst possible combination for an automation programme that is growing. It is also cloud-only with no self-hosting and no data-residency control, so a business that has promised a client its records stay on infrastructure it controls cannot use Zapier for that data at any price.
Bottom line: Zapier is unbeatable for a two-or-three-action automation between two supported apps built by a non-technical owner, and it is the most expensive place in the category to run anything longer or busier than that.
What Zapier actually costs at small-business volume
Zapier bills a task for each successful action a Zap completes. Our standing comparison workload - a lead-handling automation that fires 2,000 times a month with about six steps - is therefore around 10,000 to 12,000 billable tasks a month, depending on how many of those steps are free built-ins. Zapier's Professional plan starts at $19.99 a month billed annually ($29.99 monthly) for 750 tasks and climbs a slider to $3,389 a month annually ($5,099 monthly) at two million tasks, as of August 2026. That one automation lands well up the slider. The same job is 2,000 units on a platform that meters whole runs. Nothing about the work changed - only the unit, which is why automation billing units are worth understanding before you sign anything.
The counterweight is real and people forget it: triggers, polling checks, and the built-in Filter and Formatter steps consume no tasks, and failed actions are not billed either. That makes cheap pre-filtering a legitimate cost strategy. A Zap that triggers on every new row but filters 90 percent of them out before the first real action bills you for the 10 percent, not the 100. Designing that filter early is worth more than any plan choice.
The other line that catches people is seats, not tasks. Professional includes one user as of August 2026. The moment a second person needs to edit Zaps, you are on Team, which starts at $69 a month billed annually ($103.50 monthly) for 2,000 tasks and up to 25 users. That is a jump of roughly three and a half times the entry price, triggered by a colleague rather than by volume. Budget for the second editor from the start if you expect one.
Trigger speed is also a plan feature: 15-minute polling on Free, 2 minutes on Professional, 1 minute on Team and Enterprise as of August 2026. If your automation is customer-facing and 'instant' matters, that number is part of your price.
| Tasks/mo | Professional (annual) | Professional (monthly) | Team (annual) | Team (monthly) |
|---|---|---|---|---|
| 100 | Free | Free | n/a | n/a |
| 750 | $19.99 | $29.99 | n/a | n/a |
| 2,000 | $49.00 | $73.50 | $69.00 | $103.50 |
| 5,000 | $89.00 | $133.50 | $119.00 | $178.50 |
| 10,000 | $129.00 | $193.50 | $169.00 | $253.50 |
| 2,000,000 | $3,389.00 | $5,099.00 | $3,999.00 | $5,999.00 |
The slider, at the volumes small businesses actually hit
Zapier's headline price is the bottom rung of a ladder, and the rungs are steep. Read off zapier.com/pricing on 2026-08-06, Professional runs 750 tasks at $19.99 a month billed annually, 1,500 at $39, 2,000 at $49, 5,000 at $89 and 10,000 at $129, with monthly billing at roughly 50 percent more: $29.99, $58.50, $73.50, $133.50 and $193.50 respectively. Team begins at the 2,000-task rung - the 750 and 1,500 tiers exist only on Professional - at $69 annually or $103.50 monthly, then $119 or $178.50 at 5,000 and $169 or $253.50 at 10,000.
Two things fall out of that table. First, annual billing is not a rounding difference: it is about a third off, so a business that is confident it will still be running the automation in six months should not pay monthly. Second, the marginal cost of tasks falls as you climb, which means the worst place to sit is just over a rung. If you are consistently at 5,100 tasks, you are paying the 10,000-task price for 5,100 tasks of value, and that is the moment to either filter harder or price a different platform.
Where the money actually goes
For most small businesses the bill is not driven by the number of automations. It is driven by one busy Zap with more actions than it needs. The audit that pays for itself takes twenty minutes: open Zap history, sort by task consumption, and look at the top three. Almost always at least one of them is doing a lookup that could be a filter, or writing to two systems where one would do, or running on every record when it could run on a change.
Bottom line: Professional costs $89 a month billed annually at 5,000 tasks and $129 at 10,000, and because Professional includes exactly one seat, a second editor moves you to Team and adds roughly 30 to 40 percent at the same volume.
The pricing gotcha: held runs and the AI multiplier
When you exhaust your task allowance, Zapier does not fail your automations - it holds them. Zapier's own documentation is explicit: held runs are paused rather than lost, and can be replayed from Zap History once you have tasks available. This is genuinely more humane than dropping the data, and it is also the mechanism by which a quiet month becomes a bad week. Nothing errors. Nothing is deleted. Your automations simply stop producing outcomes until somebody notices, tops up or upgrades, and then replays the backlog by hand.
The replay itself has a speed limit worth knowing before you rely on it. Where a flood of events triggers a protective hold, Zapier documents processing replayed runs at a rate of one per second - so a thousand held runs takes roughly a quarter of an hour to drain. That is fine for a nightly sync and not fine for appointment confirmations.
The second gotcha is AI steps. Zapier's own note on task counting is that a task is a successful unit of work, and some AI actions and programmatic calls - including calls through Zapier's MCP endpoint - count as tasks in their own right, with certain AI steps consuming more than one. An AI-heavy Zap therefore costs more than its step count implies, and the gap widens exactly as you lean harder on the feature Zapier is currently promoting hardest. Model the AI steps separately from the plumbing steps.
Third, and easy to miss when comparing pages on this site: Zapier Agents is a separate product on a separate meter. It bills activities rather than tasks, so an agent's consumption does not draw down your Zap allowance and does not show up in the same forecast. Two meters is not automatically worse than one, but it does mean you cannot budget the whole account from a single number.
Bottom line: Running out of tasks does not break your Zaps, it holds them silently and replays at one run per second - which is fine for a nightly sync and useless for appointment confirmations.
What breaks in production, and how you find out
The failure mode Zapier users actually complain about is the Zap that turns itself off. Zapier's default error behaviour is to disable a Zap after repeated failures rather than keep retrying - there is a setting to keep running if errors occur, but the default is the safe one, and safe here means silent. Combine that with expired OAuth credentials, which is the most common underlying cause, and you get an automation that ran perfectly for six months and has been dead for two weeks. Nobody gets an alert unless somebody set one up.
Rate limits are published, which is more than most competitors manage, and the numbers are worth designing around. Zapier documents that Zaps with instant triggers hit 429 errors above 20,000 requests every 5 minutes per user, and that on Free and trial plans, polling triggers are held above 200 requests every 10 minutes per Zap. Private apps are limited by the owner's plan at 100 requests every 60 seconds on Free and Professional, or 5,000 every 60 seconds on Team and Enterprise, as of August 2026. Most small businesses never approach these; anyone syncing a bulk import will hit the polling one on day one.
There are hard structural ceilings too. Zapier limits a Zap to 100 steps including everything inside paths, and allows 1,000 fields per action step. During a free trial only the first 30 steps of a Zap will run, which produces a memorable class of bug where an automation works in evaluation and behaves differently the day you pay.
Finally, retention. Zapier states that it retains Zap history for 29 to 69 days, and only Enterprise admins can customise that - and customisation there means shortening it, to a range of 7 to 30 days. So your debugging window is roughly a month to two months, and your audit trail is not permanent. If you need to prove what happened in a customer dispute six months later, write your own log to a spreadsheet or a database as a step in the Zap. That is one extra task per run and it is almost always worth it.
The three-line monitoring setup Zapier does not give you
Zapier's built-in notifications tell you when a Zap errors. They do not tell you when a Zap stops being triggered, which is the failure that actually hurts, because a trigger that never fires produces no errors to report on. The cheap fix is a second Zap on a schedule trigger that checks a timestamp your main Zap writes on every run, and emails you if it is stale. Two steps, no tasks on the check itself if you use a Filter, and it catches the class of problem nothing else will.
The second habit is a monthly look at task consumption against your rung on the slider, because held runs are silent and being one task over is the same as being ten thousand over. The third is an owner: a named person who receives the alerts and is expected to act, rather than a shared inbox that nobody reads.
Bottom line: Zapier disables a Zap after repeated failures by default rather than retrying, so the most expensive failure on the platform is an expired OAuth token nobody was alerted about.
What it connects to that you already run
This is the reason to pay the premium, and it is a good reason. Zapier's catalogue is the largest in the category by a wide margin and its strength is the long tail rather than the obvious names. Google Workspace, Microsoft 365, Slack, HubSpot, Pipedrive, Salesforce, Shopify, Stripe, Square, QuickBooks, Xero, Mailchimp, Calendly, Twilio, Airtable and Notion are all covered, as you would expect. What you cannot easily get elsewhere is the practice-management system your dental clinic runs, the job-costing tool your builders use, or the booking platform specific to your trade.
If your deciding constraint is one niche piece of software that your business genuinely depends on and nothing else integrates with, that alone can justify choosing Zapier over a cheaper per-run competitor, and you should not feel bad about it. The trap is paying that premium for a catalogue you finished using in week one - if every app in your automation is a mainstream name that everything supports, you are buying breadth you are not consuming.
One caveat on the catalogue: some apps are designated premium and are gated behind paid tiers, so a plan upgrade can be forced by which app you need rather than how much you run. Check the tier of every app on your list, not just the price of the plan.
Bottom line: You are paying for the long tail, not the builder: if every app in your automation is a mainstream name that every platform supports, you are buying breadth you will never consume.
Zapier alternatives: what to pick instead, and when
Almost everything in this category is cheaper than Zapier at the same workload, which means the decision is never really about price alone - it is about what you give up. Run the shared reference workload through each and the trade is visible.
Pick Make.com instead for almost any automation with logic in it. The reference workload is $9 a month on Core against $89 on Zapier Professional billed annually, and Make handles branching, loops and array data far better. What you give up is catalogue depth in the long tail and a slightly steeper first hour. For most small businesses that have already outgrown a two-step Zap, this is the obvious move.
Pick n8n instead when you have one technical person and expect the automation programme to grow. It bills per whole execution, so the reference workload is 1,000 billable units against Zapier's 5,000, and 20 euros a month covers 2,500 of them. It is also the only mainstream option on this page you can run on your own infrastructure, which is the answer when a client contract or an insurer's questionnaire says records must stay on hardware you control.
Pick Activepieces instead if you want that same self-hosting option with a builder closer to Zapier's. A flow run costs one credit regardless of step count, the community edition is MIT-licensed and free forever, and its integrations double as MCP servers. The catalogue is much smaller, which is exactly the thing you were paying Zapier for, so this only works if you check your must-have apps first.
Pick Pipedream instead if a developer is doing the building. Compute-time billing makes short multi-step workflows cheap, code steps handle anything, and OAuth is managed for thousands of apps. No self-hosting, and its plan prices are not published, so budgeting means asking.
Two things not to confuse with alternatives. Zapier Agents is Zapier's own agent product on its own separate meter, not a cheaper way to run Zaps. And if what you want is the outcome rather than the builder, Autonoly is aimed at describing the job in plain English instead of assembling steps at all.
| Platform | Billing unit | Reference workload | Realistic monthly cost | Suits | Does better than Zapier |
|---|---|---|---|---|---|
| Zapier | 1 task per successful action | ~5,000 tasks | $89 annually, $133.50 monthly on Professional | Obscure app, no technical staff | - |
| Make | 1 credit per module action | ~5,000 credits | $9 on Core (10,000 credits included) | Non-technical builder, visual logic | Price, branching and array handling |
| n8n | 1 execution per whole run | 1,000 executions | 20 EUR (~$23) on Starter, or ~$15 hosting self-hosted | One technical owner, data control | Per-run billing, self-hosting |
| Activepieces | 1 credit per run, 2-20 per AI call | 1,000 run credits plus AI credits | $20 on Plus, or ~$15 hosting self-hosted | Open source with a familiar builder | MIT licence, per-run billing, MCP |
| Pipedream | 1 credit per 30s compute at 256MB | ~1,000 credits if runs finish inside 30s | Plan prices not published | Developers wanting managed OAuth | Code steps, cost of fast workflows |
Bottom line: The same 1,000-item inbox triage costs $89 a month on Zapier Professional, $9 on Make Core and about 20 euros on n8n Starter, so the only defensible reason to stay is a connector nobody else has.
Migration and lock-in: can you get your work out?
This is Zapier's weakest position and it deserves stating plainly. There is no export format for a Zap that another automation platform can import. You can copy a Zap within Zapier, and you can read its configuration on screen, but moving to a different vendor means a person opening each Zap, writing down the trigger, the filters, the field mappings and the actions, and rebuilding it by hand in the new tool. Five Zaps is an afternoon. Fifty Zaps is a project with a budget and a testing plan, and the field mappings are where the time actually goes.
There is no self-hosted edition and no data-residency control, so the question of where your customer records physically sit is answered by Zapier, not by you. For most small businesses that is fine. For anyone who has signed a client contract or an insurer's questionnaire promising records stay on infrastructure they control, it rules Zapier out for that data regardless of price - and no plan tier changes it.
The cheap hedge, as everywhere: maintain a plain-language description of what each Zap is supposed to achieve, next to the Zap, updated when you change it. A rebuild that starts from stated business intent takes a fraction of the time of one that starts from reverse-engineering somebody's field mappings from a screenshot.
If a migration is likely, the sequencing that works is to move the busiest Zaps first rather than the simplest. The busiest ones are where the saving is, and moving them buys you the budget and the patience for the rest. Leave the two-step Zaps between well-supported apps on Zapier indefinitely if you like - there is no prize for a clean sweep.
Bottom line: No Zap exports in a format another platform can read, so migration is a person rebuilding each one by hand - which makes moving your three busiest Zaps first, not your simplest, the sequencing that pays.
Zapier FAQ
How much does Zapier really cost once you are past the free plan?
More than the headline. As of August 2026 the Professional plan starts at $29.99 a month billed monthly, but that buys 750 tasks - and a task is each successful action, not each run. A six-step automation firing 2,000 times a month is roughly 10,000 tasks, which sits several steps up Zapier's task slider. The other jump people miss is seats: Professional includes one, so a second editor means Team, which starts at $103.50 a month billed monthly. Model tasks and seats together before you budget.
What counts as a task, and what does not?
Each successful action a Zap performs counts. Triggers do not, polling does not, and built-in steps like Filter by Zapier and Formatter by Zapier do not - which is genuinely useful, because you can filter out irrelevant records for free before spending tasks on them. Failed actions are not billed either. The important exception is that some AI actions consume multiple tasks depending on the step type and model tier, so an AI-heavy Zap costs more than the step count suggests.
What goes wrong with Zapier in production?
Silent stalling is the big one: a Zap turns itself off after repeated errors, and unless someone is watching the notification it simply stops working. Second, polling latency - on lower tiers a trigger checks periodically rather than instantly, so a 'real-time' customer notification can arrive fifteen minutes late. Third, task exhaustion mid-month, which pauses everything at once and usually happens the week something unusual drove volume up. A monthly review of Zap history and a usage alert are the cheap fixes for all three.
Who should not use Zapier?
Anyone with high volume, anyone with step-heavy workflows, and anyone with a data-residency requirement. Per-task billing means cost scales with both how often you run and how much you do per run, which is the worst combination for a growing automation programme. And because Zapier is cloud-only with no self-hosted option, a business that has contractually promised a client that records stay on its own infrastructure cannot use it for that data at all, whatever the price.
Does Zapier connect to what a small business actually runs?
Almost certainly yes, and this is the real reason to pay for it. Google Workspace, Microsoft 365, QuickBooks, Xero, Shopify, Stripe, Square, HubSpot, Pipedrive, Mailchimp, Calendly, Slack, Twilio and thousands more are supported, including a long tail of niche industry tools no other platform bothers with. If your deciding constraint is a piece of software your trade or practice depends on and nothing else integrates with, that alone can justify the premium over cheaper competitors.
How locked in are we with Zapier?
More than you would like. There is no export format that another automation platform can import, so moving means a person opening each Zap, reading the configuration and rebuilding it in the new tool. Five Zaps is an afternoon; fifty is a project with a budget. The practical defence is to keep a plain-language description of what each Zap is meant to do, maintained alongside the Zap itself, so a future rebuild starts from business intent rather than reverse-engineering someone's field mappings.
Are Zapier Agents worth using?
They are convenient rather than cheap. Zapier Agents and AI actions let you build automations that reason over a request and act across your connected apps without code, which lowers the build effort meaningfully. The cost caveat is that some AI actions consume multiple tasks, so agent-style Zaps burn the meter faster than the step count implies. Treat them as a fast way to prove an idea works, then decide whether the recurring cost justifies leaving it there or rebuilding it somewhere with a friendlier meter.
Zapier or n8n for a growing business?
Zapier if your bottleneck is connector coverage and nobody technical is available. n8n if your bottleneck is cost or control. The economics are stark: n8n Cloud at 20 euros a month covers 2,500 full workflow executions as of August 2026, regardless of how many steps each contains, while the same six-step workload on Zapier is billed as roughly ten times as many tasks. If you expect your automations to grow steps rather than just count, the crossover point arrives faster than you think.
Can you help us decide whether to stay on Zapier?
That is one of the most common questions we get, and it is usually answerable in an hour. We look at your actual task consumption, count how much of it is steps versus runs, and tell you honestly whether moving is worth the disruption - sometimes it plainly is not. That review is part of the free automation audit, and if a rebuild does make sense we can do the migration and hand it over documented, with a recurring report so you can see what the change actually saved.
Looking at alternatives? Zapier is hard to beat for the first automation and hard to justify for the tenth. Our honest guidance to small businesses is to use it deliberately: it is excellent when both apps are supported and the job is genuinely two or three actions long. On lock-in, Zaps cannot be exported in any form another platform can read, so migrating means a person rebuilding each Zap by hand - which is manageable at five Zaps and a project at fifty. Autonoly targets the jobs Zapier struggles with: scraping, file processing, and multi-step logic that goes past moving a record from A to B. If you have outgrown a simple Zap, it is worth a look before you buy the next tier up. See the Autonoly review.
