In n8n vs Make, choose n8n if you want to self-host, write code inside workflows, or run high-volume and AI-heavy automations cheaply; choose Make if you want the most polished visual builder for non-developers and never want to run a server. Make (formerly Integromat) is a cloud-only SaaS that bills per module action. n8n is fair-code software you can run on your own hardware for free, and its cloud plans bill per whole workflow run. For simple, low-volume scenarios the price gap is small. For long workflows that run thousands of times a month, it decides the question.
n8n vs Make at a glance
| n8n | Make | |
|---|---|---|
| Licence | Sustainable Use License (fair-code, source available) | Proprietary |
| Self-hosting | Yes, free Community edition | No, cloud only (on-prem agent for reaching local systems) |
| Free tier | Unlimited self-hosted; cloud has a trial | 1,000 credits a month, 15-minute minimum interval |
| Cloud entry price | Starter €20/month billed annually, 2,500 executions | Core $9/month, 10,000 credits |
| Billing unit | One execution per workflow run | One credit per module action |
| Code inside workflows | JavaScript and Python Code nodes | Limited; mostly built-in functions and formulas |
| Best for | Developers, data-heavy and AI workflows | Business users, marketing and ops teams |
Prices are list prices from each vendor's pricing page as of October 2026.
Pricing: credits vs executions
This is the difference that matters most once workflows grow.
Make charges a credit for each module action. Its own pricing page describes it this way: each module action in a scenario, like adding a Google Sheets row or fetching Gmail data, counts as one credit. The Free plan includes 1,000 credits a month and runs scenarios at most every 15 minutes. Paid plans start with Core at $9 a month, then Pro at $16 and Teams at $29 when billed monthly, each with 10,000 credits at the entry tier and a one-minute minimum interval; annual billing saves about 15%. Enterprise is custom.
n8n charges an execution for each complete workflow run, regardless of step count. n8n Cloud starts with Starter at €20 a month (2,500 executions) and Pro at €50 a month (10,000 executions), both billed annually, with unlimited users and workflows on every plan. The self-hosted Community edition has no execution limits at all.
A worked example: a scenario with eight module actions that runs 2,000 times a month.
- Make: about 8 × 2,000 = 16,000 credits, more than the 10,000 included in the entry tier of any paid plan.
- n8n Cloud: 2,000 executions, inside Starter.
- Self-hosted n8n: the cost of your server. A small VPS handles this easily.
Make's per-action model is cheaper than it looks for short, infrequent scenarios, and its entry plan costs less than n8n's. It gets expensive when scenarios loop over many items, because every iteration of every module consumes a credit.
Building workflows
Make's scenario editor is one of the best-designed automation interfaces available. Modules appear as circles on a canvas, routers split paths, iterators and aggregators handle lists, and every run can be replayed with the data each module received. Non-developers pick it up quickly, and Make's library of ready-made app modules is larger than n8n's.
n8n uses a node canvas that looks similar but leans technical. Any API is reachable through the HTTP Request node, workflows can branch, loop, merge and call sub-workflows, and Code nodes run JavaScript or Python inline when a built-in node falls short. That last point is the real divide: in Make you mostly work within the modules and formula functions it gives you, while in n8n you can drop to code at any step. With about 207,000 GitHub stars and more than 9,000 commits in the past year, n8n is also one of the most active projects in our automation category.
Self-hosting and data control
Make cannot be self-hosted. It runs on AWS in the EU and North America, so every record your scenarios touch passes through Make's infrastructure. Its on-prem agent lets cloud scenarios reach systems inside your network, such as a database or an ERP, but the orchestration stays in Make's cloud.
n8n's Community edition runs in a single Docker container with SQLite and scales to PostgreSQL and queue mode with Redis workers. Credentials, workflow data and execution logs stay on your server, which matters for regulated data and for internal systems you do not want exposed to the internet. Our n8n deploy guide covers Docker Compose, HTTPS and backups. Back up the encryption key along with the database: without it, stored credentials cannot be decrypted.
The licence is the caveat. n8n is fair-code under the Sustainable Use License, not OSI open source. Internal business use is free, but you may not sell n8n itself as a hosted service, and features such as SSO, environments and Git-based version control need a paid Business licence (€667 a month billed annually) or Enterprise. If you need a truly open-source licence, read our roundup of open-source n8n alternatives.
AI agents
Both tools now pitch AI automation. n8n has an AI Agent node with tool calling, memory and vector store integrations, and it can act as an MCP client or server. Because it self-hosts, you can point it at a local model served by Ollama and keep prompts and data entirely on your own hardware. Make offers AI agents and AI modules inside its cloud, which is convenient but means your data goes to Make and to the model provider it calls.
AI workflows also make billing differences sharper. Agents call tools in loops, and on Make each of those module actions uses credits, while on n8n the whole run is one execution.
Migrating from Make to n8n
There is no importer, so you rebuild each scenario. Most Make concepts map cleanly: scenarios become workflows, routers become IF or Switch nodes, iterators become n8n's item-based processing or Loop Over Items, and webhooks remain webhooks. Anything without a native n8n node can use the HTTP Request node with the service's API. Start with your highest-volume scenarios, since that is where the savings are.
Which should you choose?
- Choose n8n if someone on your team is comfortable with Docker or code, your workflows are long or run often, you need data to stay on your own servers, or you are building AI agents. Self-hosted, it removes per-run costs entirely.
- Choose Make if your team is non-technical, your scenarios are short and low-volume, you rely on a niche app module only Make offers, or you want no infrastructure to maintain.
- Consider Activepieces if you like Make's ease of use but want an MIT-licensed tool you can self-host. It is the closest open-source match for business users.
For more head-to-heads, see n8n vs Zapier vs Activepieces, compare live GitHub stats for n8n vs Activepieces, or browse every open-source Zapier alternative.
