Which tool migrates Airtable formulas to PostgreSQL?

Of the six compared here, only AT Migrator translates Airtable formulas into SQL. Whalesync and Airbyte copy the values, NocoDB skips formula fields, Baserow drops them, and UI Bakery does not list them as supported.

Which tool keeps Airtable linked records as foreign keys?

AT Migrator, which chooses a foreign key or a junction table from each link's cardinality. The others either do not support links, copy them as arrays of IDs, or turn every link into a many-to-many relationship.

Which tool is best if I only want an Airtable-style interface?

NocoDB or Baserow. Both give you an Airtable-style interface, and either can sit on top of a PostgreSQL database that AT Migrator built.

Which tool leaves me a schema I can build a custom application on?

AT Migrator, which creates tables and columns named after your Airtable base, with foreign keys and SQL logic. The others leave flat tables, arrays and JSONB, a join table per link, or a layout built around the tool, as Baserow's is.

Which tool is best for a live copy while I keep using Airtable?

Whalesync, which keeps Airtable and Postgres in step, or Airbyte if the destination is a warehouse for reporting.

How was this comparison made?

From each tool's documentation and, for the open-source ones, its code, as of October 2026. A cross means the tool does not do it, or does not document that it does. Each detailed comparison explains the evidence for that tool.

Comparisons

UI Bakery vs Whalesync vs Airbyte vs NocoDB vs Baserow vs AT Migrator

October 9, 20264 min read

Discuss this article with AI

ChatGPT Claude Perplexity Google AI Mode GrokView as Markdown

Opens your AI tool with a prompt about this article. It can read the live page and answer your questions.

Six tools can move Airtable data toward PostgreSQL, and they are not interchangeable. Some migrate, some sync, and some import into their own app. This page shows what they all do, where they differ, and lets you weigh the differences for your own case, so you can narrow the choice quickly and read the detailed comparison for the tools you are still weighing.

What all six do

They all move your stored data out of Airtable: records, text, numbers, and dates come across in every one of them.
Five of the six write into a database you can reach directly: only Baserow keeps the data in its own layout with no direct access on its cloud service.

Everything else is where they split, and it comes down to what Airtable calculates and how your tables relate to each other.

Where they differ

The points most tools meet come first, then the ones only a few do.

	AT Migrator	UI Bakery	Whalesync	Airbyte	NocoDB	Baserow
Writes into a database you can reach directly	Yes: Your own PostgreSQL, full access	Yes: Your own PostgreSQL	Yes: Your own Postgres, tables you create first	Yes: Your own Postgres	Yes: An external database you connect, or its own	No: Only if you self-host, none on Baserow Cloud
Gives you an interface to work with the data	No: No interface, pair it with NocoDB or Baserow	Yes: Builds internal tools and admin panels on your data	No: No interface, it only syncs	No: No interface for the data	Yes: Airtable-style interface, views and forms	Yes: Airtable-style interface, views and forms
Keeps syncing with Airtable	No: A short bridge while you move, not a permanent mirror	No: One-off copy	Yes: Two-way sync, kept live	Yes: Scheduled syncs into your destination	No: One-off import	No: One-off import
Pay once for the migration	Yes: One-time payment per migration	Yes: Free tool, nothing ongoing	No: Monthly subscription while it syncs	No: Run it yourself, or a managed plan	No: Self-host it, or per-seat cloud plans	No: Self-host it, or per-user cloud plans
Formulas become SQL	Yes: Translated into SQL expressions	No: Not on its supported-data list	No: Synced as read-only values	No: Copied as values, not logic	No: Skipped on import	No: Dropped, with their values
Rollups and lookups become SQL	Yes: Joins and aggregate queries	No: Not supported	No: One-way values only	No: Copied as arrays of values	No: Imported as NocoDB fields, not SQL	No: Dropped, with their values
Links become foreign keys or junction tables	Yes: Chosen by each link’s cardinality	No: Not supported	No: Only if you build the tables yourself	No: Arrays of record IDs	No: Every link made many-to-many	No: Kept as Baserow links, one-record limit dropped
Attachments copied to storage you control	Yes: Your own S3-compatible bucket	No: Listed as supported, destination not stated	No: Airtable links that expire	No: A string in the row, file stays in Airtable	No: Import can be switched on, destination not stated	No: Re-uploaded into Baserow’s own storage
Resumes an interrupted or large run	Yes: Skips finished tables, resumes the rest	No: No resume documented	No: Not a one-off run: it syncs continuously	No: Full refresh only, re-reads the base each run	No: No resume documented	No: Large imports have needed a manual route
Clear schema you can build a custom application on	Yes: Tables and columns named after yours, with foreign keys	No: Flat tables only, no relationships	No: You design it by hand, formulas stay static values	No: Arrays and JSONB, plus its own metadata columns	No: A join table for every link, whatever its shape	No: Tables and columns named by ID, logic in its metadata

Writes into a database you can reach directly

AT Migrator

Your own PostgreSQL, full access

UI Bakery

Your own PostgreSQL

Whalesync

Your own Postgres, tables you create first

Airbyte

Your own Postgres

NocoDB

An external database you connect, or its own

Baserow

Only if you self-host, none on Baserow Cloud

Gives you an interface to work with the data

AT Migrator

No interface, pair it with NocoDB or Baserow

UI Bakery

Builds internal tools and admin panels on your data

Whalesync

No interface, it only syncs

Airbyte

No interface for the data

NocoDB

Airtable-style interface, views and forms

Baserow

Airtable-style interface, views and forms

Keeps syncing with Airtable

AT Migrator

A short bridge while you move, not a permanent mirror

UI Bakery

One-off copy

Whalesync

Two-way sync, kept live

Airbyte

Scheduled syncs into your destination

NocoDB

One-off import

Baserow

One-off import

Pay once for the migration

AT Migrator

One-time payment per migration

UI Bakery

Free tool, nothing ongoing

Whalesync

Monthly subscription while it syncs

Airbyte

Run it yourself, or a managed plan

NocoDB

Self-host it, or per-seat cloud plans

Baserow

Self-host it, or per-user cloud plans

Formulas become SQL

AT Migrator

Translated into SQL expressions

UI Bakery

Not on its supported-data list

Whalesync

Synced as read-only values

Airbyte

Copied as values, not logic

NocoDB

Skipped on import

Baserow

Dropped, with their values

Rollups and lookups become SQL

AT Migrator

Joins and aggregate queries

UI Bakery

Not supported

Whalesync

One-way values only

Airbyte

Copied as arrays of values

NocoDB

Imported as NocoDB fields, not SQL

Baserow

Dropped, with their values

Links become foreign keys or junction tables

AT Migrator

Chosen by each link’s cardinality

UI Bakery

Not supported

Whalesync

Only if you build the tables yourself

Airbyte

Arrays of record IDs

NocoDB

Every link made many-to-many

Baserow

Kept as Baserow links, one-record limit dropped

Attachments copied to storage you control

AT Migrator

Your own S3-compatible bucket

UI Bakery

Listed as supported, destination not stated

Whalesync

Airtable links that expire

Airbyte

A string in the row, file stays in Airtable

NocoDB

Import can be switched on, destination not stated

Baserow

Re-uploaded into Baserow’s own storage

Resumes an interrupted or large run

AT Migrator

Skips finished tables, resumes the rest

UI Bakery

No resume documented

Whalesync

Not a one-off run: it syncs continuously

Airbyte

Full refresh only, re-reads the base each run

NocoDB

No resume documented

Baserow

Large imports have needed a manual route

Clear schema you can build a custom application on

AT Migrator

Tables and columns named after yours, with foreign keys

UI Bakery

Flat tables only, no relationships

Whalesync

You design it by hand, formulas stay static values

Airbyte

Arrays and JSONB, plus its own metadata columns

NocoDB

A join table for every link, whatever its shape

Baserow

Tables and columns named by ID, logic in its metadata

A check means the tool does it. A cross means it does not, or does not say that it does. Based on each tool’s documentation and, where the code is open source, the code itself, as of October 2026.

Decision matrix

Not every point matters equally to everyone. Set a weight on each one, remove the ones that do not apply to you, and see which tool fits best.

What mattersWeight: 1 nice to have, 5 must have

Writes into a database you can reach directly
You can connect to the data with any tool
Gives you an interface to work with the data
Your team can view and edit records in a UI
Keeps syncing with Airtable
Airtable stays in use and the copy stays current
Pay once for the migration
No subscription once the move is done
Formulas become SQL
Calculated fields keep working in Postgres
Rollups and lookups become SQL
Totals and pulled-in values stay live
Links become foreign keys or junction tables
Relationships keep their real shape
Attachments copied to storage you control
Files do not depend on Airtable links
Resumes an interrupted or large run
A big base can finish without starting over
Clear schema you can build a custom application on
Room to build your own app or switch tools later

Best match for you

AT MigratorBest match
80%
Meets 8 of 10 of your points
UI Bakery
30%
Meets 3 of 10 of your points
Whalesync
20%
Meets 2 of 10 of your points
Airbyte
20%
Meets 2 of 10 of your points
NocoDB
20%
Meets 2 of 10 of your points
Baserow
10%
Meets 1 of 10 of your points

Quick decision

Leaving Airtable, with formulas, links, and files that must survive: AT Migrator.
Airtable stays, and Postgres needs a live copy: Whalesync, or Airbyte for a warehouse.
An interface on top of your data: NocoDB or Baserow, or either one on a database AT Migrator built.
A small, flat base: UI Bakery's free tool.
Owning the data and building your own application later: AT Migrator. Baserow is the weakest fit.

Read the detailed comparisons

Whalesync vs AT Migrator: sync tool or migration tool.
Airbyte vs AT Migrator: pipeline or migration.
NocoDB vs AT Migrator: Airtable import or real migration.
Baserow vs AT Migrator: an app or your own PostgreSQL.
UI Bakery vs AT Migrator: a free tool for flat bases.

Ready to migrate your Airtable base?

Start a free trial on a slice of your base: 100 records, twice per account. Or scope your migration first if your setup is large, multi-base, or formula-heavy.

Start free Scope my migration