Retool

I tested Retool by building 3 internal tools from scratch on a Postgres + REST-API backend: a customer-support CRM lookup with order history, an ops dashboard for monitoring batch-job status, and an admin tool for moderating user-submitted content. The cohort I had in mind was engineering and ops teams replacing fragile internal admin tools (the spreadsheet-plus-sql-script pattern) with proper internal apps, distinct from Bubble or Webflow which target external-facing apps and from low-code form builders like Tally that target simpler use cases.

The component-and-query model is the actual product. Dragging a table component onto the canvas, wiring it to a SQL query against my Postgres warehouse, and binding row-click to a detail-edit modal took roughly 25 minutes for the support-CRM tool: that is a real engineering project that would have taken a week of bespoke React. The query library handles SQL, REST, GraphQL, and many SaaS APIs as first-class data sources, and chaining queries (run B with the result of A, conditional on user input C) works fluently once you understand the JS-expression model. The audit log captures every action for compliance, which matters for any tool that touches customer data.

What broke me was the JavaScript-everywhere learning curve and the deployment story. Retool components are powerful but customizing them past the defaults requires writing JS in narrow expression boxes that lack proper editor experience: refactoring a 60-line component-handler in a 4-line text input is genuinely worse than VSCode. The performance also degrades on apps with 30+ components and 20+ queries: page load times stretched past 3 seconds and interactions felt laggy compared to a hand-rolled React app. Pricing is per-seat with multiple add-on layers (Source Control, Workflows, AI features) and the math at 20 builders or 50 end-users gets expensive fast.

Retool is the right pick for engineering teams that need to ship 5+ internal tools per quarter where speed-to-deploy beats polished UX. Skip it if you have only 1-2 internal tools where a hand-rolled React admin is faster long-term, if your team is non-technical and needs no-code without JS expressions (Glide or Stacker are friendlier), or if your end-user count is high (50+ daily active users on internal tools) where the per-end-user cost cliff hits hard.