Your Automation Platform Is Shutting Down: Migrate One Workflow to n8n Without Rebuilding the Wrong Behavior

Freeze evidence before the source disappears

Do not begin with a blank n8n canvas. First capture the behavior you are trying to preserve:

• The source trigger and authentication boundary
• Representative valid, invalid, empty, duplicate, and wrong-source inputs
• Field names before and after every transformation
• Branching rules and stored state
• Every irreversible action, such as sending, charging, deleting, or updating
• The final receipt that proves the run completed correctly

This evidence has a short shelf life. In n8n itself, deleting a workflow also deletes its execution history, according to the official execution documentation. Treat the legacy platform's history as equally fragile: export or sanitize what you are allowed to retain before access ends. Do not copy production credentials or live customer data into a migration bundle. Use redacted samples and document the credential owner separately.

Choose one representative workflow

Migrating an entire fleet first multiplies every mistaken assumption. Choose one workflow that includes the important risk surfaces: a real trigger, at least one transformation, a branch, an external side effect, and a terminal result. Avoid both the easiest toy workflow and the most exceptional one. The first workflow is a canary. Its job is to expose the migration method's weaknesses before you repeat them across the rest of the estate.

Write a source-to-target contract

Before building, fill in a table like this:

The mapping matters because n8n expressions reference data from previous nodes and are sensitive to node structure and names. The official data-mapping guide shows how upstream output becomes downstream input. For custom code or complex fan-out, also verify item lineage. n8n's item-linking documentation explains why programmatic nodes may need explicit pairedItem information. A workflow can look correct in the editor while pairing the wrong input with an output.

Keep credentials outside the artifact

The migration artifact should contain workflow logic, placeholder credential references, sanitized fixtures, and instructions. The buyer or system owner should connect production credentials in the target environment. This is not just a hygiene preference. n8n's source-control environment guide describes credentials as stubs when moving configuration between environments. It also notes that source control works from the current saved version, not the published version. Record exactly which version you tested.

For self-hosted targets, run n8n's security audit before production handoff. It can surface credential, database, filesystem, node, and instance-level risks.

Prove five acceptance cases

A successful happy-path run is necessary, but it is not sufficient.

Store the fixture, execution identifier, terminal output, and any mutation receipt for each case. Hash the final artifacts if you need to prove that the handed-off version is the version that passed. When replaying a failed execution, document whether you used the current saved workflow or the original workflow version. n8n supports both retry modes, so an undocumented retry can test different logic from the failed run; see all executions.

Compare outcomes, not node counts

Two workflows can have different shapes and still implement the same contract. They can also look nearly identical while differing at the exact point that matters: source validation, zero-item handling, field lineage, retry behavior, or the ordering of an irreversible action. Review the evidence in this order:

• Was the event from the intended source?
• Did the required fields retain the correct meaning?
• Did every branch stop or continue for the right reason?
• Could a retry repeat an irreversible side effect?
• Did the run produce one terminal receipt?

Only then compare maintainability, node count, and execution cost.

Canary, hand off, then scale

Run the migrated workflow against sanitized fixtures first. Then use a controlled canary with the smallest safe production exposure. Keep the legacy workflow available for rollback until the acceptance evidence is reviewed. The handoff package should include:

• A source-to-target map
• The sanitized n8n workflow artifact
• Five acceptance fixtures and results
• Artifact hashes
• Credential-connection instructions
• A recovery and rollback runbook
• The exact saved or published version that was accepted

Only after this package survives review should you repeat the method across the remaining workflows.

A practical path if the deadline is close

• Free: Run the 11-check n8n reliability audit.
• $49: Get a written workflow teardown if you need the failure boundary identified before rebuilding. The full purchase price is credited toward the $299 or $799 service when upgraded within seven days.
• $299: Book the one-workflow n8n migration pilot. It covers one representative workflow, up to two external services, and the five acceptance cases above, delivered within 72 hours after complete sanitized intake. It uses the same three-slot inventory as the one-workflow reliability fix.

The faster route is not to rebuild more nodes. It is to make one workflow provably correct, then reuse the evidence-backed method.