Your Airtable base has hit the record limit

Every figure below is quoted from Airtable’s own pricing and help pages · checked 30 August 2026

This is not a bug and not something support can lift for you. It is a published property of the plan, and it has a second half that catches people out: the ceiling is per base, and the upgrade that raises it is priced per person. So the bill for more rows is set by how many colleagues can edit, which is a different question from how much data you have.

The numbers, and which one is per what

From Airtable’s own plans documentation and its pricing page:

• Free — 1,000 records per base, 1 GB of attachments per base, and five collaborators with Editor or Creator permissions. Read-only people are unlimited and commenters are capped at fifty.
• Team — 50,000 records per base, 20 GB per base, at $20 per user per month billed annually.
• Business — 125,000 records per base, 100 GB per base, at $45 per user per month billed annually.

Read those units carefully, because the two limits scale on different axes. Records are per base. Money is per editor. A team of eight that needs 2,000 rows in one base is not buying 2,000 rows; it is buying eight seats.

Splitting into more bases buys one more ceiling, not headroom

The first workaround everyone finds is to move the older rows into a second base. It works, and it is worth knowing exactly what it costs you, because the thing you give up is the thing you probably built a database for.

Airtable’s linked-record fields join tables within a base. Split a base in two and the link does not follow: the rows on the other side become text you typed, not a record you can open. Rollups and lookups that depended on that link stop being possible, views cannot span the split, and a filter that used to answer a question about all your data now answers it about half.

So splitting is a real option for genuinely separate things — last year’s archive, a different department — and a poor one for a single dataset that grew. If you find yourself splitting something you would rather keep whole, that is the ceiling doing its job, not you organising badly.

What the upgrade actually costs, as arithmetic

Because the price is per editor, the honest way to see it is a table rather than a headline. Airtable’s own annual rates, and a flat plan for comparison:

Our free plan allows unlimited apps at 10,000 entries each with 10 GB of files, and puts no count on how many people may edit. Above that it is $29 a month flat with unlimited users — there is no seat count anywhere in it, so adding the whole team, a contractor, or a client who needs read access does not move the bill.

Where Airtable’s numbers beat ours, stated here rather than left for you to find

A page like this one is easy to write dishonestly by omission, so:

• Their paid record ceilings are far above our free one. 50,000 and 125,000 per base against our free 10,000. If you need 40,000 rows, their Team plan holds it and our free plan does not — ours stops being the smaller number only on the $29 plan, where the cap is 1,000,000 entries per app.
• Their attachment allowance is per base. 1 GB each even on the free plan, so somebody with a dozen bases has more file room than our 10 GB across everything.
• At one editor, Team is cheaper than our paid plan. $20 against $29.
• Airtable is the more finished product — automations, interfaces, an extension marketplace, years of polish. We have none of that depth.

What Airtable does not publish, and we will not invent

Their plans page does not give a maximum number of bases, and we are not going to repeat a figure we cannot source. Their Enterprise Scale tier is priced on application and their pricing page mentions 100M records under it, which is a capability statement rather than a plan limit you can read off and budget against. If you have seen a specific number for either quoted elsewhere, it did not come from those two pages.

Why we wrote this

We build Bionic Forms and we would like you to consider it, so this is not neutral. What we offer instead of neutrality is that every Airtable figure here is quoted from Airtable’s own published pages rather than our characterisation, with the date we read them, and that we say where we do not fit.

Where we help: there is no editor count, so hitting a limit is never solved by removing a colleague’s access. Records live in a real database with per-person permissions and their own revision history, so “who changed this and what did it say before” has an answer. The desktop app keeps your data in an encrypted file on your own machine and works offline. Everything is exportable, and we have written down what happens to your data if we go out of business.

Where we do not: there is no automatic conversion of an Airtable base — the route in is CSV per table, and a linked-record cell arrives as the text that was in it unless the importer can match it back to a real record. Automations, interfaces and extensions have no equivalent here to import into; rebuilding those is a rebuild and you should price it as one. And we have no scripting language at all, which for some Airtable users is the end of the conversation.

Sourcing, plainly: the record, attachment and collaborator limits are quoted from Airtable’s plans documentation and the per-user prices from airtable.com/pricing, both read on 30 August 2026. Their prices and limits change; check yours. The description of what happens to a linked record when a base is split is ordinary behaviour of a link that cannot cross a base boundary, and we say so rather than dressing it as a citation. Our own figures are the ones in our pricing section — if the two ever disagree, ours is wrong.

Being straight about where we are: Bionic Forms is pre-launch with no external customers to point at. There is no iOS or Android app. The macOS, Windows and Linux desktop builds and the web app are what exist today, and the Windows build is not code-signed yet, so Windows shows a “Windows protected your PC” box the first time you run it — More info, then Run anyway. The browser app needs nothing installed.