# Airtable

> Airtable is a cloud relational database with a spreadsheet interface. Records live in tables with typed fields, tables link to each other, and the same data can be presented as a grid, kanban, calendar, gallery, form, or a purpose-built interface page. It adds automations, a REST API, an AI layer, and a growing set of agentic building tools, which is why go-to-market teams use it as the place their prospecting, enrichment, and reporting pipelines actually store state.

- Category: GTM Engineering (https://saastracker.org/categories/gtm-engineering)
- Website: https://airtable.com
- Starting price: Free for up to 5 editors and 1,000 records per base; paid plans from $20 per seat per month billed annually
- Free plan: Unlimited bases, 1,000 records per base, 1 GB attachments per base, 100 automation runs per month, up to 5 editors, and two weeks of revision history
- Free trial: 14-day trial of the Team plan, plus a permanent free plan
- Founded: 2012, HQ: San Francisco, California, United States, Ownership: Venture-backed; definitive agreement announced August 2026 to be acquired by Bending Spoons for an enterprise value of $1.285 billion, expected to close in late 2026
- Profile last reviewed: 2026-08-23
- Canonical profile: https://saastracker.org/products/airtable

## Overview

Most people meet Airtable as a nicer spreadsheet, and that framing undersells it. The underlying model is relational: a field has a type, a link field creates a real relationship between two tables, and a lookup or rollup pulls values across that relationship rather than copying them. That is the difference between a Google Sheet with a company name typed in forty times and a table where forty contacts point at one company record. For a GTM engineer, that distinction is the whole reason the tool is in the stack.

In a build-it-yourself go-to-market pipeline, the parts are usually an automation engine as the runtime, a search or scraping API for discovery, a language model for judgment, and something to hold the data between steps. Airtable is almost always that last part. A scraper writes rows, an enrichment step fills columns, a model scores the record, a human reviews the ones that look wrong in a filtered view, and an outbound tool reads the approved segment. Every one of those steps is an API call or a webhook against the same base, and the review step is the part a spreadsheet cannot do well and a Postgres table cannot do at all without someone building a front end.

The product has grown well past storage. Interface Designer produces read and write applications on top of the data with per-role permissions, so the sales team gets a clean pipeline view instead of raw grid access. Automations run trigger and action chains natively, including a JavaScript scripting step for anything the visual builder cannot express. Since 2024 the AI layer has moved fast: field-level AI that summarizes, classifies, or drafts on every row; Cobuilder, which generated a starting schema from a prompt; and Omni, a conversational builder that produces tables, interfaces, and automations from a description and now sits at the center of the company's positioning.

Two facts should shape any purchase decision made in 2026. The first is per-seat billing on editors, which means the cost of the tool tracks how many people need write access rather than how much data you store, and a small team can find itself paying more than the automation engine and the enrichment API combined. The second is ownership: on 4 August 2026 Bending Spoons agreed to acquire Airtable for an enterprise value of $1.285 billion, with closing expected late in the year subject to regulatory approval. Bending Spoons has a documented pattern of buying mature software and raising prices, so anyone building a dependency on Airtable should read its published limits carefully and keep an export path open.

## How it works

1. You create a workspace, then bases inside it. A base is the unit that matters: it holds tables, its record count is the number that hits the plan limit, and permissions, automations, and interfaces all scope to it. Tables hold records, fields carry types (text, number, single and multiple select, date, attachment, checkbox, formula, rollup, lookup, link to another record, user, barcode, button, and more), and the link field is what makes the whole thing relational rather than tabular.

2. Views are saved presentations of one table. A grid view with a filter, a sort, and a hidden-field set is the working surface for most people; kanban groups by a select field; calendar reads a date field; gallery renders attachments; form view creates a public intake form that writes rows. Views are per-user configurable on paid plans through personal views, so one analyst's filter does not disrupt everyone else's screen.

3. Interface Designer builds application pages on top of the same data: record detail layouts, filtered lists, charts, buttons that trigger automations, and forms, with element-level and role-level permissions. This is how a base becomes something you can hand to a salesperson or a client without exposing the schema, and it is the feature that keeps teams from bolting on a separate front-end tool.

4. Automations run inside the base. A trigger (record created, record matches conditions, form submitted, scheduled time, incoming webhook) fires a chain of actions (create or update a record, send an email or a Slack message, call an external API, run a script). The scripting action is a full JavaScript environment against the Airtable API, which is the escape hatch that lets a GTM engineer do the awkward transformation step without leaving the tool.

5. External systems talk to Airtable through the REST API with personal access tokens or OAuth, plus native two-way sync connectors for sources like Salesforce, Jira, Google Calendar, and other Airtable bases. The published API rate limit is 5 requests per second per base, and it does not rise with the plan, so any high-volume pipeline has to batch writes (up to 10 records per request) and queue rather than fan out concurrent calls.

## Best for

Small go-to-market and operations teams that need a shared, structured, permissioned place to hold pipeline data between automated steps, with a usable review interface on top and enough API surface to sit in the middle of a custom prospecting or enrichment workflow.

## Not the right fit for

- Datasets in the millions of rows; the per-base record ceilings (1,000 free, 50,000 Team, 125,000 Business, 500,000 Enterprise Scale) are hard limits, and performance degrades well before the cap on wide bases with many formulas.
- High-throughput integrations; 5 API requests per second per base is a fixed ceiling on every plan, including Enterprise, so it is a poor backend for anything with real write concurrency.
- Teams where most people need edit access on a budget; editor seats are the pricing axis, so ten editors on Business is $450 a month before anything else in the stack.
- Workloads that need real database guarantees: transactions, enforced referential integrity, arbitrary joins, or SQL access. Airtable's link fields are relationships, but the query surface is not SQL.
- Anyone who wants a self-hosted or fully data-resident option; Airtable is a hosted service, and regional data residency is an enterprise-level arrangement rather than a plan checkbox.

## Features

### Data model

The relational core that separates Airtable from a spreadsheet.

- **Typed fields**: Over twenty field types including select, date, attachment, user, checkbox, rating, barcode, and button, so a column enforces a shape instead of accepting any string.
- **Link to another record**: A true relationship between two tables that maintains both sides automatically, which is what turns a set of tabs into a data model.
- **Lookups and rollups**: Pull a field from a linked record, or aggregate across many of them (count, sum, max, concatenate), without duplicating the underlying value.
- **Formula fields**: A spreadsheet-like formula language evaluated per record, covering text manipulation, date math, conditionals, and regular expressions.
- **Field-level validation and defaults**: Required fields, unique constraints on primary keys, and default values that keep automated writes from producing half-formed rows.

### Views and interfaces

The presentation layer, which is where non-technical colleagues actually work.

- **Grid, kanban, calendar, gallery, timeline, and Gantt views**: Six presentations of the same records, each with its own filters, sorts, groupings, and hidden fields, saved as a named view.
- **Personal and locked views**: Personal views let one person filter without changing anyone else's screen; locked views stop a collaborator from quietly breaking the report a dashboard depends on.
- **Form view**: A shareable public form that writes directly into the table, with conditional field logic and prefilled URL parameters for campaign intake.
- **Interface Designer**: Multi-page applications built on the base: record detail layouts, filtered lists, charts, and buttons, with permissions set per element and per role.
- **Dashboards and reporting elements**: Chart, number, and pivot elements that read live from the table, which covers most internal reporting without exporting to a BI tool.

### Automation and scripting

The native runtime, useful for the steps that do not justify an external engine.

- **Trigger and action automations**: Chains started by record creation, a condition match, a form submission, a schedule, or an incoming webhook, with conditional branching and repeating groups.
- **Scripting action**: A JavaScript step running against the Airtable API inside the automation, which handles the transformation or lookup logic the visual builder cannot express.
- **Scripting extension**: Ad hoc scripts run manually against a base, the usual way a GTM engineer does a one-off backfill or a bulk deduplication pass.
- **Send request action**: A native HTTP action that calls an external API and parses the response into subsequent steps, removing the need for a middleware hop on simple integrations.
- **Automation run history**: Per-run logs with the input and output of every step, which is the difference between debugging a failure in minutes and guessing.

### AI and agents

The layer the company has rebuilt itself around since 2024.

- **Airtable AI fields**: Model-backed fields that summarize, categorize, translate, extract, or draft per record, metered from a monthly pool of AI credits attached to each seat.
- **Omni**: A conversational builder that produces tables, fields, interfaces, and automations from a plain description, then shows the generated schema and logic for editing.
- **AI in automations**: A generate-text step inside an automation chain, used for classification and drafting mid-pipeline rather than as a separate stored field.
- **Agent deployment**: Agents that operate over base data and can be surfaced in an app or a chat surface, aimed at replacing the recurring human triage step in a workflow.
- **AI credit metering**: Credits are bundled per seat (roughly 500 on Free, 15,000 on Team, 20,000 on Business) and consumed by field runs, which makes AI cost a function of headcount rather than volume.

### Integration and administration

How data gets in and out, and who is allowed to touch it.

- **REST API**: Full read and write API with personal access tokens or OAuth, batching up to 10 records per request, capped at 5 requests per second per base on every plan.
- **Webhooks API**: Subscribe to record changes and receive push notifications rather than polling, which is how an external orchestrator stays in step with the base.
- **Sync integrations**: Native connectors that mirror data from Salesforce, Jira, Zendesk, Google Calendar, Box, and other Airtable bases, with two-way sync on Business and above.
- **Extensions marketplace**: In-base apps for charts, page designer, maps, deduplication, and third-party services, plus a custom extension SDK for building your own.
- **Permissions and admin panel**: Workspace, base, table, and field-level permissions, with SSO, an admin panel, and org-wide governance available on Business and Enterprise Scale.
- **Revision history and snapshots**: Per-record change history and base snapshots, with retention that scales by plan (two weeks on Free, a year on Team and Business).

## Use cases

- **GTM engineer at a 12-person B2B startup**: Outbound targeting is assembled from three sources: a scraped list of companies that recently posted a relevant job, an enrichment API for contact details, and a model that scores fit against an ICP description. Nothing joins these together and the sales rep has no way to review the output. Outcome: One base holds Companies and Contacts as linked tables. A workflow engine writes scraped companies, an automation calls the enrichment API and an AI field scores the fit, and an Interface page shows the rep only records above a threshold with approve and reject buttons. The outbound tool reads the approved view, so the human review step exists without anyone building a UI.
- **Agency operations lead running client campaigns**: Six clients, each with their own deliverables, approval cycle, and reporting cadence, currently tracked across shared spreadsheets and a project tool that nobody updates. Outcome: Clients, Deliverables, and Assets become linked tables in one base. A form view collects briefs, rollups compute status per client automatically, and each client gets a shared Interface page showing only their own records. Weekly reporting stops being a manual assembly job.
- **RevOps analyst reconciling data the CRM cannot hold**: The CRM is the source of truth for deals, but the team also tracks partner referrals, competitor intel, and trial usage that would require custom objects and admin time to model there. Outcome: A base syncs the CRM's opportunity data one way, joins it against the tables the CRM cannot cheaply model, and pushes computed segment membership back through the API. The CRM stays clean and the messy analytical work happens where changing the schema takes a minute.
- **Founder building an internal tool without an engineer**: Inventory, supplier lead times, and reorder thresholds live in a spreadsheet that only the founder understands, and every reorder decision is a manual scan of the sheet. Outcome: Products link to Suppliers, a formula computes days of cover, an automation posts a Slack alert when cover falls below the threshold, and an Interface page gives the warehouse a simple receiving screen. No code is written and the process survives the founder being on holiday.

## Pricing

Per-editor seat subscription, billed monthly or annually, with plan tiers that gate records per base, attachment storage, automation runs, revision history depth, sync capability, AI credits, and administrative controls. Only collaborators with edit or creator permission are billed; commenters, read-only viewers, and form submitters are free. AI usage is metered in credits bundled per seat.

- **Free**: $0 per month. 1,000 records per base and 1 GB of attachments per base; Up to 5 editors, with unlimited commenters and read-only viewers; 100 automation runs per month, Interface Designer, and around 500 AI credits per editor. Genuinely usable for a single small workflow, but the 1,000-record ceiling arrives quickly on anything list-based, and no sync integrations are included.
- **Team**: $20 per seat per month, billed annually ($24 billed monthly). 50,000 records per base and 20 GB of attachments per base; 25,000 automation runs per month and one year of revision history; Standard sync integrations, extensions, and around 15,000 AI credits per seat. The tier most small teams land on. The jump from Free is steep because there is no intermediate plan.
- **Business**: $45 per seat per month, billed annually ($54 billed monthly). 125,000 records per base and 100 GB of attachments per base; 100,000 automation runs per month, two-way sync, and premium connectors; SSO, admin panel, granular interface permissions, and around 20,000 AI credits per seat. Bought mainly for SSO, admin controls, and two-way sync rather than for the record ceiling.
- **Enterprise Scale**: Custom annual contract, quoted by sales. 500,000 records per base, with HyperDB for far larger read-only datasets; Enterprise Hub, audit logs, enterprise key management, and a data loss prevention layer; Admin API, extended revision history, and higher AI credit allocations.

Billing notes:

- Seats are counted per workspace, per editor. Someone with edit rights on one base in a workspace is a billed seat for that whole workspace, which is the single most common cause of a bill being larger than expected.
- Commenters, read-only collaborators, and anyone submitting a form are free, so a base that mostly serves readers is cheap to run.
- Annual billing saves roughly 17 percent against monthly on both Team and Business as published in August 2026.
- Record limits are per base, not per workspace, so splitting a large dataset across bases is a legitimate (if inconvenient) way to stay under a ceiling.
- AI credits are bundled per seat and reset monthly; heavy AI field usage on a small team runs out of credits before a large team does, because the pool scales with headcount rather than with data volume.
- The 5 requests per second per base API limit does not increase with plan tier, so buying a higher plan does not solve a throughput problem.

Value assessment: As a database, Airtable is expensive per unit of data and cheap per unit of usability. Nobody buys 125,000 records for $45 a seat; they buy the fact that a non-technical colleague can be given a filtered, permissioned screen over live pipeline data in an afternoon, and that the same data is one REST call away from every other tool in the stack. Judged as the state layer of a homemade GTM system, a two or three seat Team plan at $40 to $60 a month is reasonable next to the automation engine and enrichment API sitting beside it. Judged as a general-purpose application backend, it is poor value: the per-base record ceilings, the fixed API rate limit, and per-editor billing all push against scale. The honest rule is that Airtable earns its price where humans look at the data, and loses to Postgres or a cheaper table store everywhere they do not.

## Strengths

- A genuinely relational model that non-technical people can build and maintain, which is rarer than the crowded no-code market suggests.
- Interface Designer removes the usual reason a small team abandons a spreadsheet workflow: there is a clean, permissioned screen for the people who should not see the schema.
- A well-documented, stable REST API and webhooks make it a first-class citizen in every automation engine, so it slots into a custom pipeline without adapters.
- Native automations with a JavaScript scripting step cover the awkward middle steps without a second vendor.
- Free viewers and commenters mean the cost tracks editors only, which suits a workflow where a few people maintain data that many people read.
- Deep template and community ecosystem, so most common GTM structures (CRM, content calendar, applicant tracker) start from a working base rather than a blank one.

## Limitations

- Per-base record ceilings are hard: 1,000 on Free, 50,000 on Team, 125,000 on Business, 500,000 on Enterprise Scale. Hitting one means splitting the base or moving off the product, not paying an overage.
- The API rate limit of 5 requests per second per base applies on every plan including Enterprise, which caps throughput regardless of spend.
- Interface and grid performance degrades noticeably on wide bases with many formula, rollup, and lookup fields well before the record limit is reached, particularly on views with heavy grouping.
- There is no SQL access and no arbitrary joins; queries are filters over one table plus its links, so genuinely analytical work has to be exported or synced elsewhere.
- The gap between Free and Team is $0 to $20 per seat with nothing in between, which forces a real decision as soon as a sixth editor or the 1,001st record appears.
- AI credits are tied to seat count, so a two-person team running AI enrichment across thousands of records exhausts the pool quickly and pays for more.
- Ownership is changing: Bending Spoons agreed in August 2026 to acquire the company, and its track record on acquired software points toward price increases, which is a planning risk for anyone building a hard dependency.

## Comparisons

- **Airtable vs Pipedream**: Different layers of the same stack rather than substitutes. Pipedream is the runtime: code-first workflow steps, thousands of connectors, and a generous execution model for developers who would rather write JavaScript or Python than click a builder. Airtable is where the results land and where a human reviews them. A common pairing is Pipedream running the enrichment and Airtable holding the record, so the choice is only exclusive if your workflow has no human review step at all, in which case Pipedream plus a real database is cheaper.
- **Airtable vs Activepieces**: Activepieces is an open-source, self-hostable automation engine, which puts it against Airtable's automations rather than against Airtable itself. If the appeal of Airtable is that its automations run natively next to the data, Activepieces answers with a runtime you control and no per-editor billing, but you still need somewhere to store state and something to render a review screen. Teams that pick Activepieces for cost reasons frequently keep Airtable anyway, because replacing Interface Designer is the hard part.
- **Airtable vs Clay**: Clay looks like Airtable and is not competing for the same job. Clay is an enrichment engine: a table where each column calls a data provider, waterfalls across sources, and spends credits to find an email or a firmographic. Airtable has no data of its own. Teams that need enriched prospect data buy Clay; teams that need a durable, permissioned home for records across many workflows buy Airtable, and plenty of stacks run Clay for discovery and push the winners into Airtable for ongoing management.
- **Airtable vs Zapier**: Zapier is the automation layer, Airtable is the data layer, and the two are so commonly paired that Airtable is one of Zapier's most used apps. Where they overlap is Airtable's native automations, which are cheaper (bundled with the plan) and faster against the base itself but limited to what happens in and around that base. Zapier wins on breadth of connectors and multi-app orchestration; Airtable's automations win on anything where the base is both the trigger and the target.
- **Airtable vs n8n**: n8n is the self-hostable, node-based engine that a technical GTM team reaches for when Zapier's per-task pricing stops making sense. Against Airtable it is again a layer difference, not a rivalry: n8n orchestrates, Airtable stores and displays. The one real trade-off is that n8n users often already run Postgres, which makes the case for paying per Airtable editor weaker, so the decision comes down to whether anyone non-technical needs to look at the data.
- **Airtable vs Make**: Make sits between Zapier and n8n on price and complexity, with a visual scenario builder and operation-based billing. Its Airtable module set is thorough, including search, upsert, and bulk operations, so it is a natural pairing when a workflow is too branching for Airtable's own automations but you do not want to self-host. As with the other engines, the question is not Make or Airtable but whether Airtable's bundled automations already cover the chain.

## Implementation

- Setup time: Minutes to a working base, an afternoon to a real one. Importing a spreadsheet and setting field types is immediate; designing the table relationships properly is the part that takes thought, and getting it wrong is the most common reason a base becomes unmanageable six months later.
- Learning curve: Low to start, moderate to do well. Anyone comfortable with a spreadsheet can use a grid view on day one. The concepts that take longer are linked records, rollups, and the discipline of one table per entity rather than one tab per topic. Automations and the scripting step assume some technical comfort.
- Onboarding: Fully self-serve, with a large template gallery, a detailed support site, and an active builder community that publishes base structures for most common use cases. Enterprise Scale customers get assisted onboarding and solution engineering; nobody below that needs it.
- Migration: Getting data in is easy: CSV import, Google Sheets import, and sync connectors handle most cases. Getting out is where to plan ahead. Records and fields export cleanly to CSV per table, but interfaces, automations, permissions, and the link relationships between tables do not export in any portable form, so a move to another platform is a rebuild rather than a migration. Given the pending change in ownership, an occasional scripted export of the tables that matter is a reasonable precaution.

## Platform, API & security

- Platforms: Web app, iOS, Android, macOS and Windows desktop apps, Slack and Microsoft Teams surfaces
- API: REST API for records, fields, tables, bases, comments, and webhooks, authenticated with personal access tokens or OAuth. Writes batch up to 10 records per request, and the rate limit is 5 requests per second per base on every plan. A webhooks API pushes change notifications, an Enterprise-only admin API covers user and workspace management, and official client libraries plus a custom extension SDK are available.
- Compliance: GDPR, CCPA, SOC 2 Type II, ISO 27001, HIPAA (Enterprise Scale, under a BAA)
- Data residency: Primarily United States hosting; regional data residency options are negotiated at the Enterprise Scale level rather than selectable on a plan.
- SSO: SAML single sign-on and SCIM user provisioning on Business and Enterprise Scale.
- Security notes: Encryption in transit and at rest, field and table level permissions, admin panel controls on Business, and enterprise key management, audit logs, and data loss prevention on Enterprise Scale. Share links are public by default when created, so link hygiene is the practical security exposure for most small teams.

## Support

- Channels: Email and in-product support, Priority support on Business, Dedicated success management on Enterprise Scale, Community forum
- Documentation: Comprehensive support site covering field types, formulas, automations, and the API, with a well-maintained API reference that includes a base-specific schema browser.
- Community: A large builder community across the official forum, a substantial consultant and template ecosystem, and long-running independent publications and podcasts dedicated to the platform, which is a meaningful advantage when debugging an unusual base structure.

## Company

- Founded: 2012
- Founders: Howie Liu, Andrew Ofstad, Emmett Nicholas
- Headquarters: San Francisco, California, United States
- Ownership: Venture-backed; definitive agreement announced August 2026 to be acquired by Bending Spoons for an enterprise value of $1.285 billion, expected to close in late 2026
- Employees: ~900 (est. 2026)
- Funding: Raised over $1.3 billion across rounds, culminating in a $735 million Series F in December 2021 at an $11.7 billion valuation.

Funding history:

- Series B (2018): $52 million. Led by CRV and Caffeinated Capital.
- Series C (2018): $100 million. Valuation of $1.1 billion.
- Series D (2020): $185 million
- Series E (2021): $270 million. Valuation of $5.77 billion.
- Series F (2021): $735 million. Valuation of $11.7 billion, the company's peak.

Timeline:

- 2012: Founded in San Francisco by Howie Liu, Andrew Ofstad, and Emmett Nicholas with the goal of making relational databases usable by non-programmers.
- 2015: Public launch after a long private beta, positioned as a spreadsheet and database hybrid.
- 2018: Blocks (later renamed Extensions) introduces in-base apps, and two large funding rounds push the valuation past $1 billion.
- 2020: Automations launch, giving bases a native trigger and action runtime with a JavaScript scripting step.
- 2021: Interface Designer becomes generally available and a $735 million Series F values the company at $11.7 billion.
- 2023: Airtable AI announced, adding model-backed fields for summarizing, categorizing, and drafting per record.
- 2024: Cobuilder generates working base schemas from a prompt and becomes the fastest-adopted feature in the company's history.
- 2025: Omni ships as a conversational app builder, and the company repositions around an AI-native platform narrative.
- 2026: Bending Spoons agrees on 4 August to acquire Airtable for an enterprise value of $1.285 billion, with closing expected late in the year subject to regulatory approval.

## Integrations

Zapier, Make, n8n, Slack, Salesforce, HubSpot, Jira, Google Workspace, Microsoft Teams, Zendesk, GitHub, Box, Dropbox, Twilio, Stripe, Google Analytics

## FAQ

### What is Airtable used for?

Airtable is used as a shared relational database with a spreadsheet interface. Teams use it to track anything with structure and relationships: CRM pipelines, content calendars, inventory, applicant tracking, campaign assets, and project work. In go-to-market engineering specifically, it is most often the storage and human-review layer sitting between an automation engine, an enrichment API, and an outbound tool.

### How much does Airtable cost?

There is a permanent free plan for up to 5 editors and 1,000 records per base. Paid plans are per editor seat: Team at $20 per seat per month billed annually ($24 monthly), Business at $45 per seat per month billed annually ($54 monthly), and a custom-quoted Enterprise Scale plan. Only editors are billed; commenters, viewers, and form submitters are free.

### Who owns Airtable?

As of August 2026 Airtable is still an independent, venture-backed company, but Bending Spoons announced a definitive agreement on 4 August 2026 to acquire it in an all-cash deal at an enterprise value of $1.285 billion. The transaction is expected to close in late 2026 subject to regulatory approval, after which Airtable would be part of Bending Spoons.

### Is Airtable a real database?

It is a real relational data store, in the sense that tables have typed fields and link fields create maintained relationships with lookups and rollups across them. It is not a real database in the sense a developer usually means: there is no SQL, no arbitrary joins, no transactions, and no enforced referential integrity. Treat it as an operational data layer with a good interface, not as an application backend.

### What is the record limit in Airtable?

Records are capped per base, not per workspace: 1,000 on Free, 50,000 on Team, 125,000 on Business, and 500,000 on Enterprise Scale. These are hard ceilings rather than metered overages. In practice, performance on wide bases with many formula and rollup fields degrades noticeably before the ceiling is reached.

### What is Airtable's API rate limit?

5 requests per second per base, and it does not increase on higher plans, including Enterprise Scale. Writes can batch up to 10 records per request, which is the main lever for throughput. Any pipeline expecting concurrent writes should queue and batch rather than fan out, or use the webhooks API to receive changes instead of polling.

### Airtable vs Google Sheets: which should a small business use?

Sheets if the data is genuinely flat, the calculations matter more than the structure, and everyone editing it understands spreadsheets. Airtable once the same entity appears in more than one place, because link fields stop the duplication that eventually breaks a sheet, and because Interface Designer and permissions let you give someone a screen instead of the raw file. Sheets is free; Airtable starts costing money at the sixth editor.

### Can Airtable replace a CRM?

For a small team with an unusual sales process, yes, and many do exactly that. What you give up is email sync, call logging, sequence tooling, native reporting on pipeline velocity, and the vendor doing schema upgrades for you. What you gain is a data model that matches how you actually sell. The common resolution is to keep a real CRM for deal records and use Airtable for everything the CRM would need custom objects to hold.

### Do free viewers count toward Airtable's seat pricing?

No. Billing counts collaborators with editor or creator permission on at least one base in the workspace. Commenters, read-only collaborators, and anyone filling in a shared form cost nothing. This makes Airtable unusually cheap for workflows where a few people maintain data that many people read.

### What is Airtable Omni?

Omni is Airtable's conversational AI builder, launched in 2025. You describe the application you want and it generates the tables, fields, interfaces, and automations, then exposes the generated schema and logic so you can edit them directly. It absorbed most of what the earlier Cobuilder feature did and is now the company's primary positioning for the product.

### Does Airtable have a free plan, and is it usable?

Yes, permanently free for up to 5 editors, with unlimited bases, 1,000 records per base, 1 GB of attachments per base, 100 automation runs per month, and two weeks of revision history. It is genuinely usable for one small workflow. The constraints that force an upgrade first are usually the automation run count and the editor limit, not the record cap.

### How do I get my data out of Airtable?

Tables export to CSV individually, and the REST API can read every record programmatically, which is the reliable route for a full backup. What does not export is the structure around the data: interfaces, automations, permissions, view configuration, and the link relationships between tables all have to be rebuilt elsewhere. Given the pending acquisition, a scripted periodic export of critical tables is a sensible habit.

## Editorial verdict

Airtable is the default answer to a specific question: where does the data live between the automated steps, and how does a human look at it. That is a narrower job than the marketing suggests, and it is a job Airtable does better than anything at its price. The relational model is real, the API is stable enough to build on, and Interface Designer solves the problem that kills most homemade workflows, which is that non-technical colleagues will not work inside a raw table. The costs are equally clear: per-editor billing that scales with headcount rather than value, hard per-base record ceilings, a 5 requests per second API limit that no amount of money removes, and no SQL when the analysis gets serious. Buy it as the state and review layer of a go-to-market stack, size the seat count honestly before committing, and keep an export script running, because the Bending Spoons acquisition announced in August 2026 makes the pricing trajectory over the next two years a genuine open question rather than a safe assumption.

---

Source: SaaSTracker (https://saastracker.org), an independent editorial project. This profile is compiled from public information, carries no peer reviews or paid placement, and was last reviewed 2026-08-23. Awards are judged on published criteria: https://saastracker.org/methodology