n8n vs Make for companies in 2026: which automation platform to choose?

The Difference in Model

Make is a proprietary cloud service. Celonis acquired Integromat in 2020, before the product became Make. Customers do not have a self-hosting option.

n8n offers a cloud service and a self-hostable Community Edition. Its source code is available under the Sustainable Use License. This is not an open-source license under the Open Source Initiative definition, as it restricts some commercial uses of n8n. A company using the product for its internal processes can nevertheless operate the Community Edition under the published terms.

Self-hosting gives the company more control over the instance and storage. It also makes the company responsible for updates, backups, secrets, logs, access rights and incidents.

Public Pricing on August 13, 2026

Prices change. The amounts below are those displayed by the vendors on the review date for the specified tiers and billing conditions.

Make

Make bills credits. Most non-AI operations consume one credit, while some features use a dynamic number of credits. Estimate a scenario from its actual modules and executions.

These figures are for Make's displayed 10,000-credit tier with annual billing. Tax and currency can change the billed amount.

n8n Cloud

n8n bills its cloud plans by complete workflow execution, with unlimited steps within one execution.

Features, execution-history retention and collaboration options also differ by plan. Volume alone is not a complete comparison.

Self-Hosted n8n

The Community Edition does not charge for each execution. It is not free to operate. Budget for the server, database, backups, monitoring, execution storage and maintenance time.

A small VPS may be enough for a light workload, but there is no defensible fixed price without knowing:

• concurrent executions;
• file and history size;
• workflow duration;
• supporting services;
• availability and recovery objectives.

My guide to a self-hosted stack with Hetzner, Coolify and n8n describes one possible architecture. It does not replace workload-based sizing.

Compare a Workflow, Not a Launch Count

A Make scenario consumes credits as operations run. An n8n Cloud workflow generally counts as one complete execution even if it contains several nodes. This can favor n8n for a long flow. Make may remain less expensive for a short or occasional flow.

Consider an illustrative scenario: a request triggers a CRM lookup, quote creation, email and record update. If four Make modules run on each pass, 300 requests represent at least 1,200 operations before iterations and retries. On n8n Cloud, 300 completed runs represent 300 executions under its billing model.

This is only a baseline. Loops, errors, sub-workflows, AI operations and testing can change consumption. Both platforms provide histories that let you measure a prototype before choosing a plan.

French SaaS Integrations and APIs

Connector lists evolve. Pennylane and Qonto should not be described as native n8n nodes unless they appear in the current official catalog.

For these services, the reference path is generally an HTTP Request node calling the vendor API:

• Pennylane documents an API for accounting and invoicing resources;
• Qonto provides a Business API with authentication methods that depend on the integration;
• n8n and Make can both send HTTP requests and process JSON.

A community connector may reduce setup work, but it introduces another maintainer and package to audit. For a financial workflow, an explicit HTTP call often makes the contract easier to inspect and API-version changes easier to track.

The article on automating Pennylane with n8n shows this approach.

Data, Security and Operations

Self-hosting n8n lets you choose the host and region. It does not mean all data stays on that server. Every called API receives the data sent by the workflow. Backups, notifications and monitoring tools can create other copies.

For either platform, document:

• the data sent to each service;
• secrets and associated permissions;
• execution-history retention;
• transfers and subprocessors;
• deletion procedures;
• who owns incident recovery.

GDPR does not automatically require self-hosting. It does require organizations to understand and control their processing. A properly governed cloud service can be appropriate. A poorly maintained self-hosted instance may be less so.

Learning and Maintainability

Make often makes simple scenarios approachable through its interface and ready-made applications. n8n becomes attractive when the team already handles webhooks, JSON, expressions or code and wants more control over its environment.

This assessment is subjective. Ask the future maintainer to build the same prototype in both tools, then check whether that person can:

• understand a failed execution;
• replay it without creating a duplicate;
• rotate an API key;
• test a change;
• trace a piece of data to its source.

The best tool is the one the team can operate after its original author leaves.

Which One Fits Your company?

Make is a reasonable candidate when scenarios are short, its connectors already cover the tools and the company does not want to run infrastructure.

n8n Cloud fits when execution-based pricing better matches the workflows, the plan has the required features and the team prefers its technical model without managing a server.

Self-hosted n8n becomes relevant when greater control over the instance is required and someone clearly owns operations. A low VPS price should not hide that work.

The most reliable decision is to prototype one representative flow on both platforms, then compare several weeks of consumption, errors, maintenance time, recovery effort and total cost.

Sources Consulted

• Make, pricing, accessed August 13, 2026.
• Make, how credits work, accessed August 13, 2026.
• Make, 2020 review and Celonis acquisition, accessed August 13, 2026.
• n8n, pricing, accessed August 13, 2026.
• n8n, Sustainable Use License, accessed August 13, 2026.
• n8n, self-hosting documentation, accessed August 13, 2026.
• n8n, integration catalog, accessed August 13, 2026.

For a prototype based on your data and constraints, Kirako offers AI agents and automation and n8n consulting for companies.