Self-Hosted n8n: Costs, Business Workflows, and Setup Compared With Zapier

Moving customer records between apps, processing invoices, and sending follow-up emails can take hours each week. Tools such as Zapier and Make, formerly Integromat, automate that work, but their subscription costs can rise as more processes depend on them.

n8n offers a self-hosted alternative. A business can run its automation platform on its own server rather than pay a hosted service for each task or execution. A virtual private server, or VPS, estimated at $20-30 per month can provide a starting point, though someone still needs to maintain and secure it.

The pricing and integration figures in this guide reflect November 5, 2025. The cost comparison is an estimate, not a quote for a particular workload.

How n8n works

n8n, pronounced “n-eight-n,” connects apps and services through workflows built in a visual drag-and-drop editor. A trigger starts a workflow, such as a payment arriving or a support ticket being created. The following steps retrieve data, apply rules, and carry out actions in other systems.

The platform is described as open-source and offers more than 400 pre-built integrations. Standard workflows don't necessarily require programming, while code nodes allow custom JavaScript when the built-in steps aren't enough. Integrations with OpenAI, Anthropic Claude, and other large language model providers add options for processing documents and generating text.

Self-hosting gives the operator control over the installation, data storage, and update schedule. It also avoids the task-based usage limits central to the cost comparison with Zapier. That doesn't mean a server has unlimited capacity: workflow complexity and execution volume still consume resources.

The estimated cost difference

Zapier's basic plans start around $20 per month in this comparison, with costs reaching $600 or more as task volume and requirements grow. Multi-step workflows and premium integrations can affect the plan needed.

Consider a growing business with these requirements:

• 15-20 active workflows.
• More than 50,000 tasks per month.
• Workflows with 10 or more steps.
• Premium integrations for CRM, accounting, and similar systems.

The estimated Zapier cost for that workload is $400-800 per month. The proposed self-hosted n8n budget is much smaller:

Those figures suggest a reduction of roughly 90% in platform and hosting costs. They aren't a full operating budget. Server administration, monitoring, backups, and security still require work, and the estimate doesn't establish that this server size will handle every workload described above.

The appeal is strongest when task volume is driving up a hosted bill and the business already has the technical capacity to run a service reliably.

What self-hosting changes

Control over data handling

Zapier processes workflow data on third-party infrastructure. With a self-hosted n8n installation, the business controls the server where the automation runs and stores its data. Connected services still matter: a workflow that sends an invoice to an AI provider or updates a hosted CRM passes data to those services.

This control can be relevant to healthcare operations with HIPAA requirements, financial services with data residency restrictions, enterprises protecting intellectual property, and international businesses handling GDPR and data sovereignty obligations. Self-hosting provides control over part of that setup; it doesn't establish compliance by itself.

Usage limits and capacity

Hosted automation plans can limit monthly tasks, available workflows, and execution time. These constraints can make a business selective about which processes it automates or complicate long-running workflows.

The self-hosted approach described here has no artificial caps on workflow counts, steps, or executions. Available server resources still constrain throughput, and increasing those resources adds cost. External services connected to a workflow also remain outside the operator's control.

Customization and maintenance

n8n supports custom integrations for proprietary systems and JavaScript code nodes for specialized logic. Access to the source code also gives technically capable operators more room to inspect behavior and address problems.

Community-maintained integrations can offer options beyond a vendor's product roadmap, but their continued maintenance isn't guaranteed. Control over updates also creates a responsibility: someone has to decide when to deploy changes and check that existing workflows still work.

Eight business workflows to build

The following examples show how a trigger can connect several business processes. Their time savings are estimates, not measured results from named deployments.

1. Customer onboarding

A new customer can require records in several systems, a welcome email, a project, and internal notifications. One workflow can coordinate those steps after a Stripe payment arrives:

1. Create a customer record in Airtable.
2. Generate a welcome email with personalized onboarding links.
3. Create a ClickUp project with templated tasks.
4. Notify the customer success team in Slack.
5. Add the customer to a Mailchimp nurture sequence.
6. Schedule a check-in reminder for seven days later.

The estimated saving is 20-30 minutes per customer. At 100 new customers per month, the upper end of that estimate is 50 hours.

2. Invoice processing and reconciliation

Invoices arriving by email need data extraction, purchase-order matching, approval, and accounting records. A workflow can begin when an email contains an invoice PDF:

1. Extract invoice data using optical character recognition or AI, such as OpenAI GPT-4 Vision.
2. Match the extracted data against a purchase order in the enterprise resource planning, or ERP, system.
3. If the records match, automatically approve and file the invoice in the accounting software.
4. If they don't match, send it to a manager for review through Slack.
5. Update a Google Sheet tracking dashboard.
6. Generate a weekly reconciliation report.

The estimated saving is 10-15 minutes per invoice. At 200 invoices per month, that could exceed 40 hours, depending on how many require manual review. Automatic approval depends on the extraction and matching logic being reliable enough for the business's approval rules.

3. Lead enrichment and qualification

Sales teams often research company details before deciding whether a lead deserves immediate attention. A new CRM lead can trigger the following process:

1. Enrich company and contact data through Clearbit or Hunter.io.
2. Check company size, revenue, and technology stack.
3. Score the lead against the ideal customer profile.
4. Add qualified leads to a high-priority sequence and other leads to a long-term nurture campaign.
5. Assign the lead to a sales representative based on territory.
6. Send a personalized introduction with relevant case studies.

The intended benefit is less research time before outreach. The qualification rules determine which prospects receive attention first.

4. Content distribution

A published blog post can trigger distribution through an RSS feed or webhook, which sends a notification when an event occurs:

1. Generate social media variations using Claude AI.
2. Schedule posts for LinkedIn, Twitter/X, and Facebook.
3. Add the post to the weekly newsletter queue in ConvertKit.
4. Update the content calendar in Notion.
5. Notify the marketing team through Slack.
6. Track URL performance in an analytics dashboard.
7. Generate a summary for an executive report.

At an estimated 45 minutes saved per post, eight posts per month would save six hours.

5. Support ticket routing

Manual triage adds work before a support agent can address the customer's problem. A workflow can begin when a help desk ticket is created:

1. Use AI to classify the ticket as billing, technical, or sales-related.
2. Check whether the customer has a free, paid, or enterprise account.
3. Route the ticket according to urgency and category.
4. Escalate enterprise tickets immediately and notify the account manager.
5. Attach relevant documentation to technical tickets.
6. Retrieve and attach account data for billing tickets.
7. Set a service-level agreement, or SLA, timer based on the customer tier.

This can shorten the path to the appropriate team. Faster responses and better customer satisfaction are intended outcomes, rather than guaranteed results of adding automation.

6. Expense reports

An expense submission through Google Forms or Typeform can replace an email-based approval chain:

1. Create an expense record in the accounting system.
2. Attach receipt images from email or cloud storage.
3. Check the expense against company policy.
4. Automatically approve expenses under $100.
5. Send expenses over $100 to a manager for approval.
6. Queue approved expenses for reimbursement.
7. Update the budget tracking spreadsheet.
8. Generate a monthly expense report.

The estimated saving is 15 minutes per report, which can add up across hundreds of monthly submissions. The example's threshold rules leave expenses of exactly $100 undefined, so that case needs a decision before deployment.

7. Inventory and orders

An order in Shopify or WooCommerce can start a workflow connecting the storefront, warehouse, shipping service, and financial reporting:

1. Check inventory in the warehouse system.
2. Create a pick list and notify the warehouse if stock is available.
3. Trigger a supplier reorder when stock is low.
4. Update the inventory database.
5. Email an order confirmation to the customer.
6. Generate a shipping label through ShipStation.
7. Add the tracking number to the order record.
8. Schedule a delivery notification for the estimated arrival.
9. Update the financial dashboard with revenue data.

The workflow reduces manual coordination across the systems involved in fulfillment. Its error handling needs to account for steps that don't complete successfully.

8. Meeting follow-up

A Google Calendar event ending can trigger follow-up work when meeting notes are available:

1. Use AI to extract action items from the notes.
2. Create tasks in Asana or ClickUp.
3. Assign tasks to the team members mentioned.
4. Send a recap email to attendees.
5. Schedule follow-up reminders.
6. Update the CRM with notes from client meetings.
7. Generate a weekly summary of completed and pending actions.

At an estimated 20 minutes saved per meeting, 20 meetings per week would save nearly seven hours.

Installing n8n

Docker

Docker provides a quick way to try n8n on a Linux VPS or a local machine that can run containers:

docker volume create n8n_data

docker run -it --rm \
  --name n8n \
  -p 5678:5678 \
  -v n8n_data:/home/node/.n8n \
  docker.n8n.io/n8nio/n8n

For a local installation, open http://localhost:5678 to access the workflow builder. This command is a starting point for trying the software, not a complete production configuration.

npm

Another installation option uses npm:

npm install n8n -g
n8n start

n8n Cloud

n8n also offers a hosted service at n8n.cloud for businesses that don't want to manage infrastructure. It charges by execution. The cost comparison presented here expects it to be less expensive than Zapier, but gives no hosted plan or equivalent workload quote to establish the difference.

Preparing a production installation

Business-critical workflows need more than a running application. A production setup should include:

• HTTPS through a reverse proxy. Nginx or Caddy can sit in front of n8n and handle encrypted connections.
• Persistent storage. SQLite is the default option for small setups. PostgreSQL and MySQL are proposed alternatives, but database support needs checking against the n8n version being deployed.
• Authentication. Enable login protection before exposing the service.
• Automated backups. Back up workflows and data regularly.
• Monitoring. Configure health checks and alerts so failures become visible.

Self-hosting also makes security an operating responsibility. Production access should use HTTPS rather than plain HTTP. Firewalls should restrict network access to the instance, and permissions should limit who can view or change workflows.

n8n encrypts stored credentials, but the encryption key still needs protection. Keep n8n and its dependencies updated, and review workflow access and permissions regularly.

Keeping workflows reliable

Start with one repetitive process rather than attempting to automate every department at once. A complete, tested workflow makes it easier to judge whether the time saved justifies the effort of maintaining it.

Add notes explaining what each step does and why it exists. Export workflows as JSON and store them in Git to track changes and support rollbacks.

Error handling should be part of the design. Failed steps need appropriate retries, and unresolved problems should notify a person who can act. Before processing 10,000 customer records, test the workflow with 10 and check both the results and any unexpected behavior.

Review execution logs after deployment. They help identify failures, slow steps, and workflows that need adjustment as connected systems change.

When a hosted service is the better fit

A business without the technical capacity to maintain a server may be better served by Zapier, Make, or n8n Cloud. Lower hosting costs don't remove the work of operating the service.

Support requirements also deserve attention. n8n offers paid support, but the appropriate comparison is the coverage and response commitment each vendor offers for the required plan. A broad assumption that one service has better enterprise support isn't enough to make the decision.

Integration coverage can be another constraint. A niche or proprietary system may require a custom integration or node, adding development and maintenance work.

AI integrations extend the available workflows to unstructured emails, documents, and images. They can generate personalized content and use context or patterns in data to inform processing. Those capabilities don't mean that a workflow automatically learns from its results; the behavior depends on the model, instructions, and workflow design.

A practical trial is to spend an hour identifying a time-consuming repetitive task, then another hour prototyping an automation. The prototype needn't be production-ready or self-hosted from the start. It should show whether the visual builder and available integrations fit the process, and where manual review remains necessary.

The n8n documentation covers installation and workflow building. The n8n community forum provides shared workflows and discussions about implementation problems.