n8n vs make.com vs Custom Code: Choosing Your Automation Stack

We use all three in production. Here's the decision framework.

n8n:

Best for: complex logic, custom code nodes, self-hosting, and teams that want control.

n8n is a visual workflow builder that lets you drop into JavaScript/Python when the no-code blocks aren't enough. Self-hosted version is free and runs on a $5/month VPS. We use it for high-volume automations (1,000+ runs/day) where we don't want per-execution pricing.

Downsides: steeper learning curve, you have to maintain the server.

make.com (formerly Integromat):

Best for: fast prototyping, non-technical teams, and citizen developers.

make.com has the best visual UX in the category. If your client's ops team needs to edit workflows themselves, make.com is the right choice. Starts free, paid plans from $9/month.

Downsides: per-operation pricing gets expensive at scale, limited when you need custom code.

Custom Python/FastAPI:

Best for: AI-heavy workflows, complex data transformations, or when you need full control over latency and reliability.

When we're building an AI pipeline that processes documents, routes to different models, stores in vector databases, and writes back to APIs — custom code is cleaner and cheaper than fighting a no-code tool.

Our decision tree:

• ▸Client team needs to edit it → make.com
• ▸High volume (1k+ runs/day) → n8n self-hosted
• ▸Involves AI/ML logic → custom code
• ▸Quick prototype for a non-tech client → make.com
• ▸Internal tool your team will maintain → n8n or custom code

When in doubt, we start with n8n. It's the easiest to hand off to a client who wants to stay hands-on.