Merge

The Merge node combines data from multiple workflow branches into a single output. Use it after parallel branches (from If-Else, For Each, or explicit splits) need to converge and continue as one path.

When to use Merge

Configuration

Merge modes

Append

Combines all inputs into a single array. Each branch's output becomes an element.

Input:

• Branch A output: { name: "Alice", source: "LinkedIn" }
• Branch B output: { name: "Bob", source: "Twitter" }

Result:

[
  { "name": "Alice", "source": "LinkedIn" },
  { "name": "Bob", "source": "Twitter" }
]

Merge by field

Combines objects by matching a shared key field.

Input:

• Branch A: { email: "alice@co.com", name: "Alice" }
• Branch B: { email: "alice@co.com", company: "Acme" }

Merge on: email

Result:

{ "email": "alice@co.com", "name": "Alice", "company": "Acme" }

First completed

Takes the output of whichever branch finishes first. Useful for race conditions or fallback patterns.

All

Passes all branch outputs as separate variables: {{merge.branch_1}}, {{merge.branch_2}}, etc.

Output variables

Common patterns

Combine If-Else outputs

Trigger → If-Else (check type) →
  Type A: HTTP Request (API 1) → Transformer (format A)
  Type B: HTTP Request (API 2) → Transformer (format B)
→ Merge (first completed) → Send Email (with unified data)

Parallel enrichment

Trigger → [parallel branches]:
  Branch 1: Company Enrichment
  Branch 2: Person Enrichment  
  Branch 3: HTTP Request (custom API)
→ Merge (merge by field: email) → Create Record (enriched contact)

Aggregate For Each results

After a For Each loop, the Merge node collects all iteration outputs:

Find All (leads) → For Each → TinyGPT (classify) → Merge (append) →
  Transformer (calculate stats: 40% enterprise, 35% mid-market, 25% SMB) →
  Send Email (classification report)