I've recently been a big fan of Claude Code for building lots of these kinds of tools, but then I always ran into the "what now" problem. Our IT team certainly wasn't going to just deploy my randomly-built tool and give it access to our systems, so there was a whole other process to go through to get the things I was building out to the team.
That's basically solved with Retool, because all the resources and data connections are already provisioned and can be centrally managed and governed. By the time I have an app built, it's already built on top of our existing permissioning structure and I can deploy it to the whole team.
With the tooling we launched today, I'll actually find myself starting in Claude Code, sketching out what we're going to be building and having CC go into my Retool account (via MCP) to find what data I actually have available, before deploying the whole thing onto Retool at the very end.
To me, that's still a flow that's going to work in 2027!
Claude is extremely good at connecting to databases. I don't see how Retool can be a big company if the moat is "you've already given us your database connection, so you might as well build and host on Retool".
For Claude, all it has to do is install a Postgres client library if your database is Postgres, etc.
In the past, we've tried using Retool before because it was tedious to build the frontend. Our backend developers wanted to build dashboards without having to learn frontend tooling.
Nowadays, our backend developers are flying with Claude Code building the frontend for them. No Retool needed. Retool would just get in the way.
The biggest problem for Retool is that you've always needed developers to build Retool but developers can move way faster building on Claude Code alone than Claude Code + Retool.
The pricing model of Retool is still absolutely crazy. How are you still able to charge per internal user per month? If I have 100 internal users, that's $18k alone. Totally crazy model in 2026.
