Make vs Zapier: Which Automation Platform Should You Pick in 2026?

Quick answer: Make vs Zapier in 60 seconds

Pick Zapier if you want the widest app catalog and automations that run with almost no setup. Pick Make if your workflows branch, loop and reshape data, and you want more control for the money. For AI agents the two are close enough that billing, not the AI, decides it. Both have free plans.

In practice the split is clean. Simple business automations, form to CRM or email to spreadsheet, ship faster on Zapier. Workflows built on routers, iterators and aggregators go further on Make. Zapier charges for completed actions; Make charges for every module that runs. That difference drives most of the cost gap.

Are Make and Zapier the same company?

No. Make is the platform that used to be called Integromat, and it took its current name on February 22, 2022. Celonis acquired Integromat in 2020. Make later launched as a business unit inside Celonis, as Munich Startup reported at the time. Zapier has no such parent. It operates as an independent company.

Confusion here is easy to explain. Both tools sit in the same category, connect many of the same business apps, and land in the same listicles. The rebrand made it worse. Years of Integromat tutorials and forum threads still circulate under the old name, so newcomers meet two names for one product and assume a corporate link. There is not.

What changed in 2025 and 2026

Make renamed its billing unit. According to Make's pricing page, operations became credits as of August 27. The company's comparison page describes a credit-based model as of November 2025. In that model each module that processes data typically consumes at least one credit, while AI or multi-bundle features may consume more. Nothing about how you build changed. What you count changed.

AI took the rest of Make's year. Make AI Agents arrived in beta, buildable with Make's own AI Provider on every plan or with your own LLM key. Maia landed as a plain-language layer over the builder. Make MCP Server exposes scenarios as callable tools for outside AI clients. Make Grid, included in all paid plans, draws real-time maps of how workflows, data flows and dependencies relate.

Zapier's changes started earlier and cut deeper into packaging. It retired the Starter plan on April 2, 2024, which made Professional its cheapest paid tier. The same repricing introduced an Enterprise plan aimed at rolling automation out to more teammates and departments with extra governance. Since then Zapier has shipped Copilot on every plan, Agents, Chatbots and its own MCP server.

Pricing for AI steps moved too. Since June 15, 2026, AI by Zapier steps are priced by model tier instead of a flat per-step rate, available on Professional, Team and Enterprise. A single step that reaches 75 tasks in one run pauses the Zap for manual approval.

Make vs Zapier at a glance

Figures marked as vendor counts come from the platforms themselves and do not reconcile. Read this as a shape, not a scoreboard.

Building automations: scenarios vs Zaps

Make gives you a canvas. Modules sit as bubbles you wire together, and the layout carries meaning. A router splits the flow into visible parallel branches, an iterator breaks an array into single bundles, an aggregator puts them back together. Make's own comparison says most scenarios can hold an unlimited number of modules, which is why long data-processing chains end up here rather than in a linear editor.

Two additions matter for debugging. As of April 2026, If-else and Merge modules were in open beta. Scenario Run Replay re-runs a workflow against historical data instead of waiting for the trigger to fire again. On paid plans, Make Grid maps how your scenarios relate, which matters once you have more of them than you can hold in your head.

Zapier works down a list: a trigger, then steps, top to bottom. Branching comes from Paths, and the documented ceilings are 3 nested Path steps, up to 10 path branches inside the same path group, and 100 steps total per Zap. That is generous for what most teams actually build and tight for what Make users enjoy building. Filters cut runs short, Formatter reshapes text, dates and numbers, and Looping repeats a set of steps.

Zapier's comparison says Make earned its power-user reputation on unlimited Routers, Iterators and Aggregators. It claims Zapier has since closed the gap with Looping, improved Paths, Sub-Zaps, Custom Actions and Code by Zapier. Both sides sell code as the escape hatch. Zapier Code steps let you drop a Python or JavaScript step inside a Zap. On Make's paid plans the Make Code app runs custom Python and JavaScript, and Enterprise plans add Custom Functions applied inline.

Failures show up in the bill more than the interface. On Zapier, failed steps do not consume tasks, so a broken run costs attention but not quota. On Make, what a failed or test run costs depends on whose account you read, covered below.

Integrations and actions per app

App counts are where the vendors stop agreeing. Zapier's pricing page claims it connects to more than 9,000 apps, which it says is more than any other automation tool. Zapier maintains those connectors and their API updates in-house. Make's own comparison says it supports 3,000+ app integrations. Zapier's comparison claims Make is closer to 3,500+ apps, roughly 900 of them community-maintained. No independent count settles it, so treat both figures as marketing arithmetic and verify your own apps are covered.

Depth per app is the more useful question. Make's own comparison says it exposes roughly double the number of preset actions per app. It cites Xero as the example: 25 actions on Zapier against 84 on Make, searches included. If your work lives inside two or three heavy systems, that difference decides more than any catalog total. If it spans thirty light tools, the catalog wins.

Neither catalog is a hard limit. Make's HTTP module will call any REST API you can authenticate against, and Webhooks by Zapier does the same on its side. Zapier lists webhooks among the Professional plan's headline features. Anything with an API is reachable from both. The only question is how much of the request you assemble by hand.

Pricing and limits in 2026: credits vs tasks

As of September 2026, Zapier's Free plan costs $0 a month. It includes 100 tasks plus Zap workflows, Tables, Forms and Copilot, with free Zaps capped at two steps: one trigger, one action. Professional starts at $19.99 a month billed annually for 750 tasks, or $29.99 billed monthly. It adds multi-step Zaps, unlimited premium apps and webhooks. Team runs $69 a month annually or $103.50 monthly for 2,000 tasks and 25 users, with shared Zaps, folders and SAML SSO. Only Enterprise offers unlimited users, and its price is a custom contract, not a published figure. The paid task slider tops out at 2 million tasks a month.

"If you're on a Professional 750 plan, your maximum task limit consists of 750 tasks from your normal task limit, plus 1500 tasks from your pay-per-task limit. This gives you a maximum task limit of 2250 tasks."

Source: help.zapier.com

Zapier advertises a 33% discount for paying yearly, and another 15% off paid plans for non-profits, excluding pay-per-task charges. Every new account starts on a 14-day Professional trial, no credit card.

Make's Free plan gives 1,000 credits a month, access to 3,000+ apps and a 15-minute floor between scheduled runs, with credits expiring at the end of the term. With yearly billing, Make lists Core at $9 a month, Pro at $16 and Teams at $29, each for 10,000 credits a month; monthly billing runs $12, $21 and $38. Enterprise is custom pricing with overage protection and advanced security features.

What counts is the real difference. Zapier defines a task as anything it successfully completes for you: sending the email, creating the spreadsheet row. Polling checks, filters and failed steps are free. Only completed actions draw down the quota, and that quota is pooled across the account, so Zaps, AI steps, code, MCP and SDK all eat the same monthly number.

Make counts modules. Each module action consumes a credit, and Make's comparison page adds that any module processing data typically consumes at least one, with AI or multi-bundle steps consuming more. Zapier's comparison claims it is worse than that: every step on Make consumes credits including polling and errors, and testing too. The accounts do not match and neither vendor is neutral, so budget with a margin. Code is metered separately at 2 credits per second.

Run one workflow through both models. Five steps, a trigger plus four actions, firing 1,000 times a month. On Zapier the trigger is free, so you spend 4,000 tasks, past what the entry paid tier includes. The same scenario on Make is five modules, roughly 5,000 credits, more if a step calls an AI model. That still sits inside the 10,000 credits that Core includes for $9 a month billed annually. The gap widens when a workflow is mostly routing and reshaping, because Zapier does not charge for steps that only decide where data goes.

Overages diverge too. According to Zapier's blog, as of June 2026, pay-per-task overage is billed at 1.25x the normal task cost on annual plans and 2.5x on monthly plans. Its help center caps total usage at three times the plan limit, so a Professional plan on its lowest tier stops at 2,250 tasks. Make sells extra credits in bundles of 1,000 or 10,000 at a fixed price set in your subscription.

AI agents, Maia and Copilot

Make AI Agents are in beta and available on every plan, built on Make's own AI Provider or on your own LLM key. You build, run and debug them inside the same Scenario Builder you use for everything else. Uploading files hands an agent context, so you skip assembling a retrieval pipeline first. Make's AI Agents page says they orchestrate processes across 3,000+ apps, and the ecosystem brings in 350+ AI apps you can wire up as ordinary modules.

Maia is the plain-language layer. As of September 2026, Make's Maia page describes it as a co-worker for AI and automation that turns written requests into automation and Make AI Agents. That same page says it is included in all plans. The vendors disagree here. Zapier's June 2026 comparison said Maia was still in early access, while Make's own page presents it as available on every tier. Make MCP Server sits alongside it, turning scenarios into callable tools for ChatGPT, Claude, Cursor and other MCP-compatible agents, covering that same catalog plus 30,000+ actions.

Zapier spreads its answer across products. Copilot ships on every plan, though Free-plan use is capped by a daily message limit that resets at 12:00 AM UTC. Zapier Agents run agent behaviors up to 400 times a month free, and Agents Pro is $400 billed annually, about $33.33 a month, for up to 1,500 activities. Chatbots Pro is $160 billed annually, about $13.33 a month, for up to five chatbots. Zapier's MCP server meters against your task pool: one tool call consumes two tasks. Zapier's June 2026 comparison frames the platform as AI orchestration, with agents acting on 9,000+ connected apps without workflows designed in advance.

Honest read: neither side has a decisive AI advantage. Make keeps agents inside the builder you already debug in. Zapier separates them into products with their own limits and price tags.

Reliability, speed, security and support

Speed is mostly a scheduling question. Make's Free plan will not run a scheduled scenario more often than every 15 minutes. Core brings that down to the minute, which Make's help center gives as the lowest regular interval. Zapier's polling follows the same shape: 15 minutes on Free, 2 minutes on Professional, 1 minute on Team and Enterprise. According to Zapier's help center, paid plans get a customizable per-Zap interval down to that minute. Webhooks are instant on both, and every Make plan includes instant webhook execution, so event-driven work is not gated behind a tier.

"Enterprise customers receive 24/7 priority support with 2-hour response SLAs, on-premises agents for securely accessing local networks, custom functions for advanced logic, comprehensive audit logs, and dedicated value engineers for implementation guidance."

Source: make.com

Uptime is asymmetric. Make Enterprise plans carry a guaranteed 99.5% uptime SLA. Zapier publishes no comparable first-party figure, so its status page is the reference point rather than a contract. Check the vendor's page if a number has to go into procurement.

Security baselines sit close together. Make's standard tier covers GDPR, SOC 2 Type II and ISO 27001, with AES-256 encryption at rest and TLS 1.2 or 1.3 in transit. According to Make's pricing page, the infrastructure runs on AWS in the EU and North America across all tiers. Make Enterprise adds SSO via OAuth2 or SAML2, role-based access controls, audit logs, on-premises agents, custom functions, dedicated value engineers and 24/7 priority support with 2-hour response SLAs.

Zapier's side reads much the same on paper. Make's comparison says Zapier Enterprise covers similar ground with SSO, advanced admin controls, Premier support, custom retention policies and activity logs. Zapier's comparison claims its connection event logs begin on Team while Make's audit logs begin on Enterprise, and lists SOC 2 Type II, SOC 3 and GDPR for both.

Zapier includes Live Chat from the Professional plan's 2,000-task tier and above. Make's own comparison says Make Academy carries 36 courses against Zapier Academy's 12, which is a count, not a quality measure.

Who should choose Make, and who should choose Zapier

A small business running a handful of simple automations should take Zapier. The linear editor is faster to learn. The catalog is likelier to hold the niche tool you signed up for last week, and the entry paid tier covers what most small teams generate. You pay more per unit of work at scale, and that is a problem for later.

E-commerce and marketing teams with real branching logic should take Make. Order routing by region, lead scoring with several outcomes, feeds that need splitting, enriching and recombining: that is what routers, iterators and aggregators exist for. Rebuilding it as nested Paths is unpleasant.

Developers and technical teams usually land on Make too, for the canvas, the module count and the code app. The MCP server is the tiebreaker if you want scenarios callable from an AI client. If your work is mostly moving records between SaaS products, Zapier's code steps will not hold you back.

For enterprise governance, read the contract rather than the feature list. Zapier is the only side offering unlimited users, and its top tier is where admin controls and activity logs live. Make Enterprise is where the uptime commitment, audit logs, on-premises agents and a named response time sit. Which wins depends on whether your blocker is seat count or a signed guarantee.

Switching from Zapier to Make, or back: a short checklist

• Recount your busiest workflows the way the target platform bills them: completed actions on Zapier, modules on Make.
• Confirm each app you actually use exists on the other side, down to the specific actions, not just the connector.
• Check whether your triggers are webhooks or polling; a polling-heavy setup shifts cost and latency more than the app list does.
• Rebuild one high-volume workflow first and run it in parallel for a full billing cycle before canceling anything.
• Re-create error handling and notifications deliberately; they rarely survive a migration untouched.
• Export or screenshot the existing logic before you switch workflows off; you will need it when the rebuild disagrees.

Make vs Zapier vs Latenode

Both platforms here bill by countable units of work: Zapier by completed action, Make by module. That is easy to audit and easy to inflate, because adding a step adds to the bill whether or not the step does anything expensive. A third model bills by how long the work takes, worth knowing before you commit to per-step pricing for years.

Latenode is built on that model. You pay for workflow runtime measured in CPU seconds, runs multiplied by average duration, with no per-operation and no per-step fees. Adding steps does not multiply the bill. The first 10,000 CPU seconds each month are free on both plans. The free plan costs $0 forever with no credit card, covering 5 active workflows, one execution worker, 3 minutes per run and a 3-day run history. It has a visual drag-and-drop builder, an AI Copilot inside every workflow and an AI agent builder.

Estimation is the honest catch. Runtime billing rewards lean workflows and punishes slow ones, but you cannot price a workflow until you have run it, a harder mental model than counting steps on a diagram. Latenode is not the winner here: if you need the biggest connector catalog or a signed enterprise SLA, the choice comes down to Zapier and Make. Runtime billing is a reason to look, not to switch on its own.