n8n vs Make in 2026: which one fits a LATAM startup (with real numbers)

May 16, 20267 minAI, Automation, n8n, Make, Comparison

Short answer (60 seconds): for a LATAM SaaS startup in 2026, Make (USD 9-29/month) is the fastest option if nobody on your team wants to maintain infrastructure. n8n self-hosted on a USD 5-30/month VPS is 4-6x cheaper above 5K-10K executions/month, but requires someone comfortable with Docker. n8n cloud (USD 24-480/month) is the worst of both worlds: paying for service without control. For AI integration, both are current with OpenAI/Anthropic; the real decision is operational, not technical.

This comparison exists because most of the ones on Google are written from Madrid with EUR pricing and give recommendations that don't apply to a founder in Buenos Aires, México DF, or Bogotá. Here numbers are in USD, latencies are measured from LATAM, and the reference use case is a SaaS startup, not a marketing agency.

Quick price table (USD, May 2026)

Critical note: Make charges by operations (each step in a scenario counts as one op), n8n by executions (each time the workflow fires). A typical execution uses 5 to 15 ops depending on complexity. That means 10K Make ops ≈ 1K-2K n8n executions in comparable terms.

Real cost per execution, with a concrete case

Use a realistic case: an agent that classifies incoming support tickets and routes them to Slack by category.

Workflow steps:

1. Trigger: incoming webhook (1 op).
2. Call OpenAI to classify (1 op).
3. Conditional decision (1 op).
4. Create ticket in HubSpot (1 op).
5. Notify the relevant Slack channel (1 op).

= 5 ops per execution in Make, 1 execution in n8n.

At 500 tickets/day (15,000/month):

At this volume Make is cheaper than n8n cloud. But scale to 100K tickets/month:

At scale, self-hosted n8n becomes nearly free compared to any cloud option. That's the real wedge.

Who wins on each dimension

UX for non-developers: Make

Make has the most polished UX in the automation market. Drag-and-drop, output preview at each step, visual error handling. If your ops team isn't technical, you'll lose fewer human-hours paying than saving on infra.

Cost at scale: n8n self-hosted

Above 5K executions/month, n8n self-hosted wins by a lot. A Hetzner CX22 (USD 5/month) runs n8n with unlimited executions until you saturate CPU/memory. For most small SaaS, that server handles 100K+ executions/month without breaking a sweat.

Control and portability: n8n

n8n is open source (Sustainable Use License — free for internal use, paid for reselling). Workflows are exportable JSON, you can version them in git, do code review. Make locks you into the platform — exporting is hard and workflows aren't portable.

Native integrations: Make slightly ahead

Make has ~1,800 native integrations vs ~1,000 in n8n at time of writing. But "native" in Make often means "the Make team built this node"; in n8n it means "someone from the community built this node". For the big ones (HubSpot, Stripe, Slack, OpenAI, Google Workspace) both are current.

Performance / latency from LATAM: n8n self-hosted in a local VPS

If you run n8n on a VPS in São Paulo (DigitalOcean) or Santiago, base latency is 20-50ms. Make runs in EU/US data centers — latency 100-250ms from LATAM. For backend automations the difference isn't noticeable; for synchronous flows where the user waits, yes.

Compliance / data control: n8n self-hosted

If your compliance requires that data NOT leave your infra, self-hosted is the only real option. Make runs on their servers; n8n cloud too. Only self-hosted lets you say "data never touches a third party".

When to pick each

Pick Make if:

• You have nobody on the team comfortable with Docker/Linux.
• Volume under 5K executions/month.
• Your #1 priority is letting any ops person build flows.
• You need very specific integrations (Salesforce, Microsoft Dynamics, Quickbase) where Make typically has better coverage.

Pick n8n cloud if:

• Your team is technical but doesn't want to maintain infra.
• You want flexibility for custom code (function nodes in JavaScript).
• Medium volume (5K-50K executions/month) and the USD 60-480 vs self-hosted gap is acceptable.

Pick n8n self-hosted if:

• High or fast-growing volume (10K+ executions/month).
• Your team has at least one dev comfortable with Docker.
• You want to version workflows in git and do code review.
• Compliance or data privacy demands it.
• You'd hate it if a vendor changed pricing 3x overnight.

Use neither if:

• The flow is complex (nested loops, dynamic conditions, more than 3-level branches).
• It'll live inside your SaaS product as a feature.
• Volume >1M executions/month (writing it in code is cheaper).
• You need fine-grained debugging — no-code debugging is slow.

Minimum n8n self-hosted setup on Hetzner

To put numbers behind "self-hosted" — here's what it costs and takes:

1. Create Hetzner CX22 VPS (2 vCPU, 4GB RAM, 40GB SSD) — USD 5.39/month.
2. Point a domain + Cloudflare for free SSL — USD 0.
3. Install n8n with Docker Compose:

~
# docker-compose.yml
services:
n8n:
  image: n8nio/n8n:latest
  ports:
    - "5678:5678"
  environment:
    - N8N_HOST=n8n.your-domain.com
    - WEBHOOK_URL=https://n8n.your-domain.com
    - N8N_PROTOCOL=https
    - N8N_BASIC_AUTH_ACTIVE=true
    - N8N_BASIC_AUTH_USER=admin
    - N8N_BASIC_AUTH_PASSWORD=${N8N_PASSWORD}
    - DB_TYPE=postgresdb
    - DB_POSTGRESDB_HOST=postgres
    - DB_POSTGRESDB_USER=n8n
    - DB_POSTGRESDB_PASSWORD=${POSTGRES_PASSWORD}
  volumes:
    - n8n_data:/home/node/.n8n
  depends_on:
    - postgres
postgres:
  image: postgres:16
  environment:
    POSTGRES_USER: n8n
    POSTGRES_PASSWORD: ${POSTGRES_PASSWORD}
    POSTGRES_DB: n8n
  volumes:
    - postgres_data:/var/lib/postgresql/data
volumes:
n8n_data:
postgres_data:

4. Automated backup to S3 / R2 with a cron + pg_dump — USD 1-2/month in storage.

Total: ~USD 7/month, unlimited executions until CPU saturates.

Setup time: 2-3 hours the first time, 30 min if you've done it before.

My honest recommendation for a LATAM SaaS startup

If I have to give a general rule:

• 0-5K executions/month, no dev: Make Core (USD 9/month). Ready in 1 hour.
• 5K-50K executions/month, with a dev: n8n self-hosted on Hetzner (USD 7-15/month). Best cost-benefit on the market.
• Above 50K executions/month: n8n self-hosted, consider Redis for queues and a bigger VPS.
• If automation is part of the product you charge for: write it in code (TypeScript with Inngest or BullMQ). No no-code platform scales to multi-tenant SaaS without pain.

Let's talk about your case

If you're choosing between n8n, Make, or building your own orchestrator, and want to review your specific case before committing, book a 30-minute call at no cost. 20 minutes usually clarifies whether the decision is trivial or has nuances that flip the answer.

Read also:

• Integrating OpenAI without blowing up costs — if your automation will call LLMs.
• How much does it cost to implement AI in a SaaS startup — full project budget.
• More articles on automation — guides and comparisons.
• Back to the blog — all articles.