n8n vs Make: credits vs executions, priced on five real workloads
Make is the cheaper hosted choice for short, low-volume scenarios: 1,000 free credits a month, and Core from $12 a month for 10,000. n8n wins when a run has many steps, loops over items or runs code, because n8n bills one execution per run and Make bills a credit for nearly every module, for every item. We priced five real workloads on both using the official pricing and credit rules on Oct 25, 2026. Watch Make's polling: a trigger that checks every 15 minutes uses 2,880 credits a month even when nothing arrives.
Verdict
- Choose Make for visual scenarios with routers and filters, a few thousand module runs a month, and no wish to host anything.
- Choose n8n for loops over many items, 10+ step workflows, AI agents with tools, code steps, or self-hosting.
- For complex workflow automation the bill decides it: Make cost scales with modules × items, while n8n Cloud cost scales with runs.
n8n vs Make at a glance
| n8n | Make | |
|---|---|---|
| Billing unit | Execution: one full run, any number of steps or items | Credit: 1 per module operation, per bundle (item). Routers, filters and error handlers are free |
| Since when | Executions on all cloud plans | Credits replaced operations on Aug 27, 2025 |
| Free option | Self-hosted Community edition; cloud trial with 1,000 executions | Free plan: 1,000 credits a month, 15 min minimum interval |
| Cheapest paid plan | Starter $20/mo annual for 2,500 executions | Core $12/mo annual ($16 monthly) for 10,000 credits |
| Polling trigger cost | Counts only when new data is found (per docs) | 1 credit per check, data or not |
| Code | JavaScript Code node, no extra cost | Make Code app: 2 credits per second |
| AI | AI Agent node on every plan, your own model key | Make AI Agents: 1 credit per operation; Make's AI Provider adds token credits; your own key needs a paid plan |
| Apps | Hundreds of native nodes plus HTTP Request | 3,500+ apps |
| When you run out | Executions cap for the month | Scenarios stop; webhooks queue. Unused credits expire at the end of the term |
| Hosting | Cloud (EU) or your own server | Make cloud only |
| Learning curve | Steeper, more developer-shaped | Visual canvas, easier for most people |
Pricing and limits (checked Oct 25, 2026)
| Make credits a month | Core (annual) | Pro (annual) | Teams (annual) |
|---|---|---|---|
| 1,000 | Free | ||
| 10,000 | $12 ($16 monthly) | $21 ($28 monthly) | $38 ($51 monthly) |
| 20,000 | $21 | $38 | $69 |
| 40,000 | $38 | $69 | $129 |
| 80,000 | $72 | $118 | $225 |
| 150,000 | $129 | $199 | $379 |
| n8n plan | Annual | Monthly | Executions a month | Key limits |
|---|---|---|---|---|
| Community (self-hosted) | $0 | $0 | Unlimited | You run the server, updates and backups; Sustainable Use License |
| Starter | $20 | $24 | 2,500 | 5 concurrent runs, 5 min max per run, 1 shared project |
| Pro | $50 | $60 | 10,000 | Up to 50 concurrent runs, 40 min max per run |
| Business | $800 | $960 | 40,000 | Self-hosted with SSO, Git environments, queue mode |
| Enterprise | Custom | Custom | Custom | Sales-led |
Prices and limits read from n8n.io/pricing, zapier.com/pricing and make.com/en/pricing on Oct 25, 2026, in US dollars with annual billing unless noted. Plans change often, so re-check before you buy. We have no affiliate deal with any of the three.
Credits vs executions: five workloads priced
| Workload (per month) | Make credits | Make plan | n8n executions | n8n Cloud plan | Why |
|---|---|---|---|---|---|
| 3-step recipe, instant trigger, 300 runs | 900 | $0 Free plan | 300 | $20 Starter, 2,500 executions | |
| Same recipe, trigger polls every 15 min, 300 runs | 3,480 | $12 Core, 10,000-credit tier | 300 | $20 Starter, 2,500 executions | Make bills each check; n8n counts a poll only when it finds data |
| 3-step recipe, 3,000 runs | 9,000 | $12 Core, 10,000-credit tier | 3,000 | $50 Pro, 10,000 executions | |
| Daily digest: 50 items, 2 steps per item | 3,090 | $12 Core, 10,000-credit tier | 30 | $20 Starter, 2,500 executions | Make bills each module per item; n8n bills one execution per run |
| 10-step workflow, 10,000 runs | 100,000 | $129 Core, 150,000-credit tier | 10,000 | $50 Pro, 10,000 executions |
Credit rules from Make's credit usage help page: triggers use 1 credit per run even with no data; action modules use 1 credit per bundle; an iterator uses 1 credit and each module after it uses 1 per item. The digest assumes a trigger, a search that returns 50 items, an iterator, then an AI step and a write for each item (103 credits a run). The polling row assumes a 30-day month. Execution rules from n8n's executions docs. Model costs (OpenAI, Claude, Gemini) are extra on both unless you use Make's AI Provider, which bills tokens as credits.
Note the first row. A 300-run, 3-step scenario fits Make Free with an instant trigger, but the same scenario on a 15-minute poll needs Core, because the checks alone use 2,880 credits. Where you can, use a webhook or an app's instant trigger. n8n has its own version of this question: our n8n pricing guide shows how schedules, webhooks and polling triggers count toward executions.
Where n8n wins
- Loops and batches. Fifty items through two steps is still one execution.
- Long workflows. The average public n8n template uses 9.1 distinct node types, and 30.7% use 11 or more.
- AI agents. 4,247 public n8n templates use an agent node. 29.0% of those add memory and 38.6% add a structured output parser, all inside one execution.
- Self-hosting with no execution cap.
Where Make wins
- Price at low volume. Core is $12 a month for 10,000 credits, cheaper than n8n Starter for most short scenarios.
- Visual clarity. Routers and filters on the canvas are easier to read than a long n8n graph, and they're free.
- Nothing to host or upgrade. n8n major versions have removed nodes, while Make updates apps for you.
- Error handlers are free modules. On n8n you build an Error Trigger workflow yourself, and only 2.7% of public n8n templates have one.
Start with a recipe
These recipes loop or poll, so they show the billing difference most clearly. Each has an importable n8n file:
- Monday digest of what's due this weekn8n JSON + Zapier blueprint
- Weekly link roundup from your saved bookmarksn8n JSON + Zapier blueprint
- Alert when a supplier changes a pricen8n JSON + Zapier blueprint
- Weekly list of newsletters you never openn8n JSON + Zapier blueprint
- Order emails to a Google Sheet with Geminin8n JSON + Zapier blueprint
- Find the questions your FAQ doesn't answern8n JSON + Zapier blueprint
See all 25 on the free n8n templates page, or compare n8n vs Zapier.
Related comparisons
- n8n vs ZapierCost of the same workflow at 100 to 10,000 runs a month
- n8n vs Zapier vs Make for AI agentsWhat one agent run costs on each
- Activepieces vs n8nCredits vs executions, plans and app coverage
- 25 free n8n templatesImport one and price it on your own plan
- n8n node libraryUsage data and known problems for the 22 most used nodes
Questions
- Is Make or n8n better?
- Make is better for visual, low-volume scenarios between common apps, and for people who don't want to host anything. n8n is better for long or looping workflows, AI agents with tools, code-heavy steps and self-hosting. If two people on the same team argue about it, the one who will maintain it should pick.
- Is it worth switching from Make to n8n for AI workflows?
- If your AI scenarios loop over many items or call several tools per run, often yes, because Make bills credits per module per item and n8n bills one execution per run. For a handful of short AI scenarios, Make Core at $12 a month is hard to beat. Remember there's no importer: each scenario is a rebuild.
- Is Make or n8n cheaper?
- For short, low-volume scenarios, Make: its Free plan has 1,000 credits and Core is $12 a month (annual) for 10,000. n8n Cloud starts at $20 for 2,500 executions. n8n gets cheaper as steps and items per run grow. In our table, a 10-step workflow run 10,000 times a month needs Make Core 150k ($129) against n8n Pro ($50). Self-hosted n8n has no licence fee.
- Does Make use credits when there's no new data?
- Yes. Make's help center says a trigger module uses 1 credit per run, whether or not it returns data. Checking every 15 minutes is 2,880 credits a month, almost three times the Free plan's 1,000. Instant (webhook) triggers avoid this. n8n's docs say polling triggers count an execution only when they find new data, but schedule triggers count every run. One forum thread reports empty Gmail polls being counted on recent versions, so watch your usage page.
- Why use n8n over Make?
- Per-run pricing that ignores step count, a free self-hosted edition, JavaScript Code nodes that don't burn extra credits, and an AI Agent node on every plan. Make's Code app uses 2 credits per second of runtime.
- Can I import Make scenarios into n8n?
- No. The formats are different and neither tool imports the other. Rebuild the scenario node by node. Routers become IF or Switch nodes, and iterators usually go away because n8n processes each item automatically.