I've been trialing the new Airtable interfaces for a few client-facing dashboards. On the surface, it's a slick way to avoid building a separate presentation layer. But let's talk about what it's actually costing, because the move from "database" to "application front-end" feels like a classic vendor lock-in play.
The pricing seems to jump from "manageable" to "enterprise SaaS" real fast when you need:
- More than a handful of interfaces per base
- Custom permissions for external users (clients)
- Any meaningful automation between the interface and the underlying data
My immediate questions for anyone using this in production:
* What's your actual seat count vs. viewer count? Is the "commenter" tier sufficient for clients, or are you forced into "editor" licenses just so they can interact with a button?
* Have you done a break-even vs. a traditional stack? For example, the cost of Airtable Pro/Enterprise seats for your team *plus* external users, compared to a Retool/Internal tool front-end on a cheaper database (Postgres on a managed instance, anyone?).
* How granular are the permission controls *really*? If a client can only see their own records via an interface, is that a true data boundary or just a view filter?
The promise is "no-code client portals." The reality, I suspect, is a monthly bill that scales linearly with your client base, not your usage. I'd love to see some real numbers from teams who've migrated.
Show me the bill
Spot on about the cost jump. We hit that wall at about five interfaces. The commenter tier felt like a trap - clients couldn't actually submit forms or update records via buttons, which defeated the whole purpose. We ended up needing editor licenses for every external user, which blew the budget.
I ran a quick ROI analysis against a Retool+Supabase setup. The breakeven was about 14 months. After that, the custom stack was cheaper, but you're paying upfront in dev time. For short-term or fast-moving projects, Airtable's interface speed might still justify the premium.
The permission controls are decent *within* an interface view, but configuring them across multiple interfaces for different clients on one base gets messy fast. You're right to be skeptical about vendor lock-in too - once your client workflows live in those interfaces, migration becomes a huge pain.
Keep automating!