CyberFreezeDev- Home
- Blog
- How-to guides
- How to migrate from Zapier to n8n without breaking anything
How to migrate from Zapier to n8n without breaking anything
Zaps do not import into n8n, so a migration is an inventory, a rebuild and a careful switch. Here is what maps directly, what has to be rebuilt, and when it is better to stay on Zapier.
To migrate from Zapier to n8n, list every Zap with its trigger, steps and monthly runs, rebuild each one as an n8n workflow, re-create the app credentials, and run old and new side by side on real inputs before you switch the trigger over and turn the Zap off. Linear Zaps map almost step for step. Paths, Formatter and Zapier's built-in tools such as Storage, Digest and Tables have to be rebuilt with n8n's own nodes, because Zaps do not import into n8n.
If you are still deciding whether to move at all, start with our comparison of n8n vs Zapier.
Step 1: take an inventory of your Zaps
Zapier will give you most of the raw material. In Settings, Security and data, Data management, choose Export my data: Zapier emails a ZIP whose Zapfile.json holds every undeleted Zap, with the data needed to run each one. On Team and Enterprise plans you can also use Download my Zap workflows for a JSON file of Zaps alone, and owners can download everyone's. From the Zap history page you can export up to 5,000 runs at a time as CSV or JSON, which tells you what each Zap really does in a month.
Put each Zap on one row of a sheet:
- Name, owner and what it is for
- Trigger app and event
- Every step, including Filters, Paths and Formatter steps
- Runs a month, from the history
- Whether it carries personal, financial or health data
- What breaks, and who notices, if it stops
Then sort. Move the Zaps with the most steps and runs first, because that is where Zapier's per-task billing costs most, and the ones carrying sensitive data you would rather keep on your own server. Short, rare Zaps can wait or stay.
What maps directly, and what needs rebuilding?
| In Zapier | In n8n | Effort |
|---|---|---|
| App trigger and action steps | The same app's trigger and action nodes, or HTTP Request | Direct, if a node exists |
| Webhooks by Zapier | Webhook node | Direct, new URL |
| Filter by Zapier | Filter or If node | Direct |
| Paths by Zapier | If or Switch node, one branch per path | Rebuild the logic |
| Formatter by Zapier | Expressions, Date & Time, Edit Fields (Set), Code | Rebuild each transform |
| Delay by Zapier | Wait node | Direct |
| Looping by Zapier | Usually nothing: n8n runs each node once per item | Simpler |
| Sub-Zap by Zapier | Execute Sub-workflow node | Direct |
| Code by Zapier | Code node (JavaScript or Python) | Port and retest |
| Storage, Digest, Tables | Data tables, or your own database | Rebuild |
| Forms, Interfaces | n8n Form, or a form on your site | Rebuild |
Formatter is where most hidden work sits. A single Zap may split a name, reformat a date, look up a value in a table and change a phone number to international format. Each of those becomes an expression or a small Code step, and each needs a test case. Check every app as well: if a niche tool has a Zapier app but no n8n node, you will be calling its API with the HTTP Request node.
How do credentials move over?
They do not. Zapier connections stay in Zapier, so every app is connected again in n8n, and that is a good moment to tidy up:
- Make a list of every connected account, the person who owns it and the permissions it needs.
- Create new API keys for n8n instead of copying old ones, so you can revoke the Zapier keys cleanly later.
- Use a shared service account, not a personal login, for anything business-critical.
- On self-hosted n8n, Google sign-ins need your own OAuth app in the Google Cloud Console. n8n says its managed Google OAuth is available on n8n Cloud only, and for certain Google nodes.
Test in parallel, then cut over
- Build on a test instance with sample data, and name the workflow after the Zap it replaces.
- Run both on the same inputs. Where the source app allows two webhook URLs, send events to both. Point n8n's writes at a test sheet or sandbox, and make sure only the Zap sends anything to customers.
- Compare results against the Zap history for one to two weeks of normal traffic, including the odd cases: a blank field, a refund, a name with an apostrophe.
- Switch over. Publish the n8n workflow, which registers its production webhook URL, update the source app to that URL, and turn the Zap off the same hour. Turn the Zap off, do not delete it.
- Watch the Executions tab and your error workflow daily for the first weeks. Keep the old Zap for a month, then delete it and revoke its credentials.
When should you stay on Zapier?
Stay when the Zaps are short, run a few hundred times a month and nobody technical will own a server. Stay when a tool your team depends on has a Zapier app and no n8n node, and its API is awkward. And stay for the parts of your setup built on Zapier's own products, such as Tables and Interfaces, unless you are ready to replace them. A mixed setup is normal: move the long, high-volume and sensitive workflows, and leave the rest.
Plan the move with us
Our n8n integrations service covers the whole route: the inventory, a test instance, the rebuild, the parallel run and the cutover, with workflows exported to Git and an error workflow watching every run. See our n8n integration page for the systems we connect. If you have a Zapier bill that keeps rising, send us your list of Zaps and we will tell you which are worth moving. Book a consultation.
