Migration guide

How to migrate from Zapier to n8n

The short answer

Migrating from Zapier to n8n means rebuilding each Zap as an n8n workflow, because there is no automatic converter. The safe order is inventory, prioritize by cost and risk, rebuild, run both in parallel, cut over one workflow at a time, then turn off the Zaps. A program of 5–15 workflows typically takes 4–10 weeks with Lumeor.

ConceptZapiern8n
ZapA Zap, made of one trigger and one or more stepsA workflow, made of a trigger node and connected nodes
Action stepAn app action, such as Create Contact in HubSpotAn app node operation, or an HTTP Request node for any REST API
Billing unitTask, counted for each successful action stepExecution on n8n Cloud, counted once per workflow run; no execution fees when self-hosted
Filter by ZapierStops the Zap unless conditions matchIf node, or the Filter node to drop items
Paths by ZapierBranches the Zap into conditional pathsSwitch node, or chained If nodes
Formatter by ZapierText, number and date transformsExpressions inside any field, the Edit Fields (Set) node, or the Code node
Code by ZapierJavaScript or Python step with limits on runtimeCode node running JavaScript or Python
Looping by ZapierRuns later steps once per line itemNative, since n8n nodes process every item they receive; Loop Over Items for batching
Delay by ZapierDelay For, Delay Until, Delay After QueueWait node
Sub-ZapsCall a Sub-Zap and Return from a Sub-Zap stepsExecute Sub-workflow node calling a workflow that starts with the Execute Sub-workflow trigger
Webhooks by ZapierCatch Hook URL on hooks.zapier.comWebhook node with its own test and production URLs on your n8n domain
Storage and Zapier TablesKey-value storage and hosted tablesData Table node, or a real database such as Postgres
Errors and replayZap history, error alerts and AutoreplayExecutions list, Retry On Fail per node, and an error workflow that starts with the Error Trigger node

Stay on Zapier if

  • You run a modest number of simple Zaps and the task bill is not a concern
  • Non-technical teams build and fix their own automations and want to keep doing so
  • Your workflows depend on niche apps that have a Zapier connector but no n8n node and a poor public API
  • Nobody on the team, or at a partner, will own n8n day to day

Move to n8n if

  • Multi-step Zaps at high volume are driving up your task count
  • Your logic needs loops, merges, sub-workflows or real code
  • You are adding AI agents and want control over models, tools, memory and prompts
  • Data must stay on infrastructure you control, which self-hosted n8n allows
  • You have, or are hiring, a technical owner for core automations

Why teams migrate

Most Zapier to n8n migrations start with the bill. Zapier counts a task for every successful action step, so a 12-step Zap that runs 8,000 times a month is up to 96,000 tasks. Zapier does not count triggers, Filter, Paths, Formatter or Delay steps, so check your real usage page before you estimate savings. On n8n Cloud the same run is one execution, and self-hosted n8n has no execution fees at all.

The second reason is complexity. Zaps that need nested branches, loops over line items, merges of two data sources or retries with custom logic become hard to read and harder to debug. n8n shows the data flowing out of every node, which makes those workflows easier to own.

The third is AI. Teams that want an agent to qualify leads, classify tickets or read documents want to choose the model, see each tool call and control cost. n8n’s AI Agent node gives you that. If you are still deciding between the platforms, read our n8n vs Zapier comparison first.

The migration plan

  1. Inventory. Export your Zap data and pull the task usage report. For each Zap, record the trigger, apps, step count, monthly runs, owner, and whether it touches money, customers or regulated data.
  2. Prioritize. Rank by task cost and business risk. Move high-volume, low-risk Zaps first to prove the platform, then the critical ones. Leave simple, rarely used Zaps on Zapier unless you are cancelling it entirely.
  3. Set up n8n properly. Decide Cloud or self-hosted (see n8n Cloud vs self-hosted). Configure credentials, an error workflow, execution data retention and access control before you rebuild anything.
  4. Rebuild, do not transliterate. Three chained Zaps passing data through Storage often become one n8n workflow. Formatter steps usually disappear into expressions. Use the mapping table above as a checklist.
  5. Run in parallel. Run the n8n workflow against the same inputs with its final write steps disabled or pointed at a test record, and compare outputs to Zapier for at least one full business cycle (a week, or a month-end for finance flows).
  6. Cut over. Switch one workflow at a time. Change the webhook URL or enable the n8n trigger, then turn off the matching Zap in the same change window.
  7. Decommission. Keep the Zaps turned off, not deleted, for 30 days. Export run history you need for audits, then delete the Zaps and downgrade or cancel the plan.

What to watch for

  • Webhook URLs change. Every Catch Hook URL on hooks.zapier.com must be replaced with an n8n Webhook node production URL in the sending system (forms, Stripe, your app). n8n’s test URL only works while you are listening in the editor, so never paste it into a live system.
  • Credentials must be re-authorized. OAuth connections cannot be transferred. Every app is connected again in n8n, which means finding the account owner for each one. Use service accounts where the app supports them so a workflow does not break when someone leaves.
  • Duplicates and gaps during cutover. Zapier deduplicates polling triggers by record ID. A new n8n polling trigger will not know what Zapier already processed. Pick a cutover timestamp, and for anything that creates records or sends messages, check a unique key (order ID, email, invoice number) before writing.
  • In-flight runs. Zaps paused in a Delay step will still finish after you turn the trigger off. Let them drain or handle them by hand.
  • Error handling is your job now. Zapier sends error alerts, and paid plans offer Autoreplay for temporary failures. In n8n, set Retry On Fail on flaky API nodes, route failures to an error workflow that posts to Slack or email, and decide which errors should stop the workflow.
  • Hidden logic. Look for Zaps that rely on Zapier-specific behavior, such as line-item handling or a Formatter lookup table, and rebuild those explicitly.

Timeline and cost

Moving a single workflow, including rebuild, testing and a parallel run, fits our single workflow automation band of $2,500–$12,000 and typically takes 1–3 weeks. Migrating a portfolio of 5–15 workflows is an automation program, typically $15,000–$50,000 over 4–10 weeks. The main cost drivers are the number of distinct apps (each needs credentials and testing), how many workflows touch money or customer data, and whether you self-host. Our n8n automations suite case study shows the kind of messaging and intelligence workflows we run on n8n, and the n8n program cost breakdown explains where the budget goes.

What we recommend

Move the expensive, complex and AI-driven work to n8n and let small team Zaps stay on Zapier. Do it workflow by workflow with a parallel run, not as a weekend cutover. If you want a second opinion on which Zaps are worth moving, start here.

Frequently asked questions

Can I import my Zaps into n8n automatically?

No. Zapier lets you export your Zap data from its account settings, which is useful for building an inventory, but n8n cannot import that format. Each Zap is rebuilt by hand as an n8n workflow. Simple Zaps take an hour or two each.

Will my Zap history move to n8n?

No. Run history stays in Zapier. If you need it for audits, export the Zap run history you need before you cancel the plan, and start early on accounts with a long history.

Is self-hosted n8n free?

The community edition is free to self-host for internal business use under n8n's Sustainable Use License. You still pay for the server, backups and someone's time to keep it patched. Some features, such as SSO and Git-based source control, require a paid license.

Do I need to move every Zap?

No. We usually move the core, high-volume and AI-heavy processes and leave small personal or departmental Zaps where they are. Running both platforms is common and sensible.

How long does a Zapier to n8n migration take?

A single workflow usually takes 1–3 weeks including parallel running. A program of 5–15 workflows usually takes 4–10 weeks, mostly driven by how many apps need new credentials and how long you want to run both systems side by side.

Next steps

Last updated October 14, 2026. Product details change often; check each vendor's current documentation and pricing.