Airtable is an AI-native app platform where teams describe an app in plain language and get working tables, interfaces, automations, and AI agents on top of their own operational data, with admin governance and enterprise compliance.
Used this tool? Rate it
Used this tool? Rate it
Airtable is an AI-native application platform built on top of a relational database that ordinary business users can operate. Teams use it to hold the operational data they actually run on — campaigns, product roadmaps, applicant pipelines, inventory, content calendars — and then build working software directly on that data without hiring engineers. Where a spreadsheet stops at rows and formulas, Airtable adds typed fields, linked records between tables, permissioned interfaces for different audiences, and automations that move work forward on their own.
The important thing to understand in 2026 is that AI is no longer a feature bolted onto that database. Airtable's own homepage leads with the claim that AI transformed individual work and that Airtable transforms teamwork, positioning itself as the system of record for companies building what's next. That repositioning was deliberate: in the June 2025 launch letter, the company wrote that instead of just adding more AI capabilities to its existing platform, it treated the moment as a refounding of the company. Independent technology press covered the same shift as a genuine platform relaunch rather than a marketing refresh.
What that means practically is that the fastest path into Airtable is now conversational. You describe the app you want, and Omni — the platform's conversational builder — assembles the tables, interfaces, and automations for you. The database is still underneath, and you can still edit every piece by hand, but the starting point has moved from "design a schema" to "explain the problem."
Airtable offers a free plan aimed at individuals, very small teams, and lightweight needs. It is genuinely usable but tightly bounded: 1,000 records per base, 1GB of attachment storage per base, two weeks of revision and snapshot history, and 1,000 API calls per workspace per month. For evaluating whether the model fits your process it is sufficient; for running a real department it usually is not.
Paid tiers are billed per user. Team costs $20 per user per month billed annually, raising limits to 50,000 records and 20GB of attachments per base with a year of revision history. Business costs $45 per user per month billed annually, with 125,000 records and 100GB per base, unlimited API calls, and the administrative controls most organizations need. Enterprise Scale pricing is custom, based on the organization's needs and scale, and is where governance features such as enterprise key management, data loss prevention, and EU data residency live.
AI usage is metered separately through credits, and this is the part buyers most often miss. Each plan includes a monthly credit allowance per paid user, and credits are pooled for shared use across all workspace or organization users rather than reserved individually. Consumption is variable — each action consumes a number of credits based on its input data, selected model, and output data — so a document-analysis agent running across a large table behaves very differently from an occasional summarization. When the allowance runs out, self-serve customers can buy add-on packs at $20 per month for 10,000 credits, $200 per month for 100,000, and $800 per month for 400,000. One useful exception: building apps and agents with Omni is free and does not consume AI credits, so experimenting with app generation does not draw down the same budget as running agents in production.
Yes, there is a free plan, and it is functional rather than a time-limited trial. It allows 1,000 records per base, 1GB of attachment storage per base, two weeks of revision and snapshot history, and 1,000 API calls per workspace per month. That is enough to evaluate the platform or run a small personal workflow, but most teams reach the record limit fairly quickly once they log ongoing operational data.
Team is $20 per user per month when billed annually and Business is $45 per user per month billed annually, with Enterprise Scale priced custom based on the organization's needs and scale. Because billing is per seat, the practical cost depends on how many people need edit access rather than on how much you use the product — which is the single most common source of unexpected bills.
Both. Each plan includes a monthly AI credit allowance per paid user, and those credits are pooled across the workspace or organization rather than reserved per person. Consumption is variable: each action's cost depends on its input data, the selected model, and the output produced. If the allowance runs out, self-serve customers can buy additional packs — 10,000 credits for $20 per month, 100,000 for $200, or 400,000 for $800. Notably, building apps and agents with Omni does not consume credits; the credits are for running AI actions.
No. Airtable states that model providers never retain your data and that your data is never used for model training. Enterprise customers additionally decide which models are enabled, choosing between providers including OpenAI, Anthropic, and Meta, which gives organizations control over where AI processing happens rather than accepting a single fixed vendor.
Omni builds complete apps — including tables, interfaces, and automations — from a plain-language description of what you need. It works from your existing data and produces components you can edit with the normal visual tools afterwards. It is best understood as a way to get a working first version quickly; you still review the schema, correct field types, and validate the automation logic before the app carries real work.
It depends on the plan: 1,000 per base on Free, 50,000 on Team, and 125,000 on Business. These are per-base limits, not per-workspace, so splitting data across bases is a common workaround. For genuinely large datasets, HyperDB syncs and scales datasets even with 100M+ records, but that is aimed at enterprise use rather than being an extension of standard plan limits.
Airtable has undergone a SOC 2 Type 2 audit and holds ISO/IEC 27001:2022 for information security management and ISO/IEC 27701:2019 for privacy information management, alongside HIPAA and TX-RAMP Level 2. Customers who need regional storage have the option to have their data stored in the European Union, and enterprise customers can store data at rest with customer-owned keys. Administrators get audit logs, granular permissions, and data loss prevention.
Residents of the EEA, the UK, and Switzerland have the right of access, portability, rectification, erasure, restriction, withdrawal of consent, and objection. Residents of applicable US states have rights to access, deletion, correction, and appeal, and the policy states you have the right to be free from discrimination for exercising your privacy rights. Airtable states it stores personal information for no longer than necessary for the purposes for which it was collected.
Not operationally, at least for now. Bending Spoons announced a definitive agreement on 4 August 2026 to acquire Airtable at an enterprise value of $1.285 billion, but the deal has not closed: it remains subject to regulatory approvals and other customary closing conditions, is expected to close later this year, and both companies continue to operate independently until then. The service, brand, and domain are unchanged. It is still worth factoring into long-term procurement decisions, since pricing and roadmap authority may shift after closing.
Yes. The Web API closely follows REST semantics, uses JSON to encode objects, and relies on standard HTTP codes to signal operation outcomes, with call quotas that vary by plan — 1,000 calls per workspace per month on Free, 100,000 on Team, and unlimited on Business. Automations also connect natively to tools such as Slack and Google Workspace, and JavaScript steps let you extend automation logic when the visual builder is not enough.
