How to Migrate from Make to n8n in 2026 (Cost Math + Step-by-Step)
You suspect you’re overpaying for Make.com. Your scenarios are growing, your credit consumption climbs every month, and the per-module billing model means complexity is penalised directly. The suspicion is reasonable: Make charges per action taken inside a scenario, so every extra module — every filter, every router branch, every iterator — adds to the bill. n8n bills per workflow execution instead, meaning one full run costs the same whether it contains three nodes or thirty. That structural difference is the entire thesis of this guide.
E-E-A-T: This guide is based on official documentation, pricing pages, and community reports — we did not run the tool hands-on. Pricing verified 2026-10-08.
The billing-unit difference, in plain words
Make and n8n both charge for automation runs, but they count different things. Understanding this distinction is the first step in any migration decision — and it explains why two platforms can produce wildly different bills for the same workload.
Make’s billing unit is the credit (renamed from “operations” in its pricing overhaul). Each module action inside a scenario costs one credit. A 10-module scenario running 1,000 times per month consumes 10,000 credits — one credit per module per run. n8n bills per workflow execution. One full run of a workflow — regardless of how many nodes it contains — counts as a single execution. A 10-module n8n workflow running 1,000 times consumes 1,000 executions, not 10,000. That is the core asymmetry: Make’s costs scale with scenario complexity; n8n’s don’t. Make pricing n8n pricing
The cost math, worked out
Now let’s put concrete numbers behind the billing-unit difference. The worked example below uses a realistic scenario shape — 10 modules, 1,000 monthly runs — and compares what Make and n8n would each charge for identical workload volume.
Worked example: A 10-module Make scenario running 1,000 times per month consumes 10,000 credits — exactly the ceiling of Make Core at $12/mo (Make pricing). The same workload in n8n consumes 1,000 executions (one per run, node count irrelevant), fitting comfortably within n8n Cloud Starter at €20/mo billed annually (€24 monthly) with 2,500 executions — a 2.5× headroom buffer (n8n pricing). Self-hosted n8n Community Edition is free software with unlimited executions; the real cost is a VPS at roughly $5–30/mo (n8n pricing). The thesis holds: Make’s credits scale with module count; n8n’s executions don’t.
| Plan | Make | n8n Cloud | Self-hosted n8n |
|---|---|---|---|
| Free / entry | $0 — 1,000 credits/mo, 15-min minimum interval | — | Free software; ~$5–30/mo VPS |
| Tier 1 | Core $12/mo — 10k credits | Starter €20/mo annual (€24 monthly) — 2,500 executions | ~$6–12/mo typical VPS |
| Tier 2 | Pro $21/mo — 10k credits + priority execution | Pro €50/€60 — 10,000 executions | — |
| Tier 3 | Teams $38/mo — 10k credits + team roles | Business €667/€800 — 40,000 executions + self-hosting licence | — |
| Billing unit | Per module action (1 credit per module) | Per workflow execution (1 execution per run) | Unlimited executions; infra cost only |
| Source | make.com/en/pricing | n8n.io/pricing | n8n.io/pricing |
Currency note: Make lists in USD, n8n in EUR. No conversion applied; the relative gap is the signal. Self-hosted costs assume a standard VPS — see our self-hosting guide for setup detail.
Step-by-step migration path
There is no import tool that converts Make scenarios into n8n workflows automatically. Migration is a manual rebuild process, but it is structured and repeatable. These eight steps cover the full path from audit to cutover.
- Inventory your Make scenarios. List every active scenario, its module count, run frequency, and credit consumption. Identify which ones are stable versus frequently edited.
- Export Make blueprints as JSON. Use Make’s blueprint export feature for reference only — these files are not importable into n8n, but they document the logic you need to rebuild.
- Map Make modules to n8n nodes. Use the mapping table below to translate each Make building block into its n8n equivalent. Most scenarios will use a small subset of these mappings.
- Recreate credentials in n8n. API keys, OAuth tokens, and connection details must be re-entered. n8n stores credentials separately from workflows, so set them up once and reference them across workflows.
- Repoint webhook URLs in external systems. Any system sending data to a Make webhook must be updated to call the new n8n webhook URL. This is the step most likely to cause downtime if missed.
- Run both platforms in parallel for 1–2 weeks. Trigger both Make and n8n workflows on the same inputs and compare outputs. Do not cancel Make until this phase completes.
- Cut over to n8n. Once outputs match, switch the primary execution path to n8n. Keep Make running in read-only or paused mode as a fallback.
- Cancel Make last. Only after n8n has proven stable in production for at least one full billing cycle should you cancel your Make subscription.
Module-to-node mapping
| Make module | n8n equivalent | Notes |
|---|---|---|
| Router | Switch node | Routes data down branches based on conditions |
| Iterator | Split In Batches (Loop Over Items) | Iterates over arrays; similar batching behaviour |
| Aggregator | Code node / Item Lists | May require a Code node for complex aggregation logic |
| Filter | IF node | Condition-based pass/block within a workflow |
| Scheduler | Schedule Trigger | Built-in trigger; no separate module needed |
| Set variable | Set node / Code node | Custom variables in Make Pro map to Set nodes in n8n |
| HTTP request | HTTP Request node | Direct equivalent; both platforms support custom API calls |
| Error router | Error Trigger / IF error handling | n8n has dedicated Error workflows; configure per-workflow |
What breaks in translation
The mapping table covers the happy path, but real-world scenarios surface friction in three specific areas. Knowing these in advance prevents mid-migration surprises.
Expression syntax differences. Make uses its own {{1.field}} token syntax with built-in functions like formatDate() and add(). n8n uses JavaScript expressions in its Expression editor — {{ $json.field }} with standard JavaScript methods. Simple field references translate directly, but complex expressions need rewriting. Date formatting, string manipulation, and math functions all use different APIs.
Make data stores vs n8n equivalents. Make’s Data Stores provide a built-in key-value store accessible across scenarios. n8n has no direct equivalent in its Cloud plans. The workaround is to use an external database (PostgreSQL, Redis, or even a simple VPS-hosted store) accessed via n8n’s database nodes. For simple use cases, the Item Lists node can hold structured data within a single workflow, but it does not persist across runs.
Error handlers. Make attaches error handlers at the scenario level with built-in routes for ignoring, resuming, or breaking on errors. n8n requires explicit Error Trigger workflows or IF-node error branching. The logic is equivalent, but the configuration is more manual — you wire error handling into each workflow rather than declaring it at the platform level.
When you should NOT migrate
Not every Make user should move. n8n has its own lock-in surfaces, and the cost savings can disappear under specific conditions. This section is the honest counterweight.
Fair-code license limits. n8n runs on the Sustainable Use License — a fair-code model, not OSI open source (n8n license). You may self-host for internal business use without paying n8n for a licence, but you may not offer n8n as a service to your customers. If your business involves reselling automation infrastructure or embedding it in a customer-facing product, the licence restricts you in ways Make’s terms do not.
Self-host maintenance burden. The $5–30/mo VPS figure covers the server, not your time. Self-hosting means managing Docker containers, applying updates, monitoring uptime, handling SSL certificates, and troubleshooting when a workflow fails at 2 a.m. If you lack server administration skills or don’t want to acquire them, n8n Cloud’s managed infrastructure is worth the €20/mo even if the per-execution math looks less favourable.
Make-only app integrations. Make advertises more than 2,000 apps in its library, many with pre-built modules that n8n lacks natively (Make pricing). If your workflows depend on niche apps available only on Make, migrating means replacing those steps with custom HTTP request nodes — possible, but fragile and maintenance-heavy. Audit your app dependencies before committing.
The verdict: who wins at what scale
The decision rule is straightforward. If your average scenario uses more than five modules and your monthly run volume is moderate, Make’s credit billing will cost more than n8n’s execution billing at equivalent scale. If your scenarios are simple — two or three modules — and your volume is low, Make’s Free tier at 1,000 credits/mo may already be sufficient, and migration effort isn’t justified.
For teams running complex, multi-module workflows at scale, the savings compound: Make Pro at $21/mo covers 10k credits; a 15-module scenario running 700 times/month hits that ceiling. The same workload on n8n Starter is 700 executions — well under the 2,500-execution limit. For a broader migration framework, see our Zapier-to-n8n migration guide — the structural logic transfers across platforms.
FAQ
Does n8n have an import tool for Make scenarios?
No. n8n provides no native import or conversion tool for Make blueprints. Export your scenarios as JSON from Make for reference, then rebuild each workflow visually in n8n’s editor. The process is manual but straightforward for most standard module types.
Is n8n actually free if I self-host?
The software is free — n8n Community Edition carries no licence fee for internal business use under the Sustainable Use License. But you pay for a VPS (roughly $5–30/mo), and you assume responsibility for hosting, security, updates, and uptime. “Free” means no licensing cost, not zero cost.
How long should I run both platforms in parallel?
One to two weeks minimum. Run both Make and n8n on identical inputs and compare outputs field by field. Extend the parallel run if your workflows involve monthly cycles, seasonal triggers, or infrequent error paths that need surface time. Do not cancel Make before this phase ends.
📖 Related Reads
- ToolBrain — tool reviews, LLM comparisons, and AI workflow guides
Cross-links automatically generated from NoCode Insider.