n8n-nodes-baserow-plus

Production-grade n8n community node for Baserow — the open-source Airtable alternative.

Built by Expert Local

Field-type-aware serialization, true batch API (up to 200 rows per request), upsert, lookup, automatic retry with exponential backoff, and parallel multi-table fetch.

Installation

Docker / Custom Nodes Directory

Copy the entire nodes-basedrow-v3 folder into your n8n custom nodes directory:

scp -r nodes-basedrow-v3/ user@server:/path/to/n8n/custom/

Then restart n8n:

docker restart n8n n8n-worker

npm (when published)

cd ~/.n8n
npm install n8n-nodes-baserow-plus

Credentials

Create a new credential of type Baserow Plus API with:

The credential name (baserowPlusApi) is distinct from v2, so both can be installed side-by-side.

Operations

List

List rows with optional filters, sorting, full-text search, and server-side field filtering.

Get

Retrieve one or more rows by ID. Accepts a single ID or comma-separated IDs.

Create

Create a new row. Supports both auto-mapping from input JSON and manual field configuration.

Update

Update an existing row by ID.

Delete

Delete a row by ID.

Batch Create

Create multiple rows using Baserow's batch API endpoint. Sends up to 200 rows per request, automatically chunking larger sets.

Performance: A single batch API call replaces 200 individual HTTP requests. Creating 500 rows requires only 3 API calls instead of 500.

Batch Update

Update multiple rows in a single batch API call. Two modes:

1. Shared fields mode: Provide comma-separated Row IDs + field values applied to every row
2. Per-item mode: Leave Row IDs empty, enable Auto Map All, and include an id field in each input item

Batch Delete

Delete multiple rows in a single batch API call. Provide comma-separated Row IDs.

Upsert

Update if a matching row exists, otherwise create. Select a field to match on — the node searches for an existing row with that field value and updates it, or creates a new row if no match is found.

Lookup

Search for rows by field value. A standalone search operation that exposes what upsert does internally.

Multi-Table Fetch

Fetch data from multiple Baserow tables in parallel using a single shared filter field. All tables are queried concurrently via Promise.all.

Field-Type-Aware Serialization

The node fetches each table's field schema and uses field type information for precise serialization:

When field type info is unavailable (e.g. cache miss), falls back to a heuristic that preserves backward compatibility with v2.

Error Handling & Retry

• Structured errors: Extracts Baserow error codes and detail messages for clear diagnostics
• 429 retry: Automatic retry on rate limiting with exponential backoff (1s, 2s, 4s). Respects Retry-After header.
• Configurable timeout: Default 30s, adjustable per-node via the "Request Timeout" parameter
• Continue on fail: All operations respect n8n's continueOnFail() pattern

Bug Fixes from v2

1. [null] → [0] in sanitizeWriteValue — Number(null)===0 passed integer check. Fixed with null/undefined guard.
2. Empty array [] → "[]" — Now passes through as [] for link_row fields.
3. Multi-select vs linked record overlap — Both checked value[0]?.value. Now checks for .color property (select options have {id, value, color}, linked records have {id, value} only).
4. Upsert matchValue broken with autoMapAll — Convoluted originalFields lookup failed. Now reads directly from items[i].json[upsertField].
5. minFilterValue rejects 0 — Used !== '' which rejected 0. Now uses != null.
6. batchCreate not batched — Was identical to create (1 API call per row). Now uses Baserow's real batch endpoint.

Configuration

License

MIT

Author

Steven Lefebvre — expertlocal.ca