An honest comparison
Custom automation
vs Retool
Retool solves a real problem well: internal tools are expensive to build and nobody wants to spend a sprint on an admin screen. The question is never whether Retool is good. It is whether the thing you are building is an internal screen or the operation itself. Those two look identical on day one and completely different in year two.
Where Retool is the right call
An admin panel for a support team. A screen that lets operations correct records without touching the database directly. A quick internal view of data that already lives somewhere sensible. In all three, the value is in the speed of getting a usable interface in front of colleagues, and the cost of the licence is small next to a week of developer time. If that is your case, use it, and do not let anyone talk you into a build.
Where it stops being a bargain
Per-user pricing turns the tool into a tax on growth: the more of your team relies on it, the more it costs, indefinitely. Logic drifts into the tool, because it is easier to add a query in the builder than to change the service behind it, and after a year the tool contains rules that exist nowhere else. And the moment someone outside your company needs a screen, the model breaks entirely, because it is built for internal users by design.
The question that settles it
Ask what happens if the tool disappears tomorrow. If the answer is that a handful of colleagues lose a convenient screen and use the database directly for a week, keep it. If the answer is that your operation stops, then the tool is not a tool, it is your system, and it should be built as one, in your repository and your accounts, with tests and monitoring and a way to know when something breaks.
The comparison, straight
| Retool | Custom system | |
|---|---|---|
| Best for | Internal admin screens over existing data | Operations your business depends on |
| Time to first result | Hours to days | 2 to 4 weeks for a first scope |
| Cost model | Per user, grows with the team | Built once, then nothing to rent |
| External users | Not the design intent | Customer portals with real permissions |
| Business logic | Drifts into the builder over time | Lives in versioned, tested code |
| Testing | Manual, against live data | Automated suites, staged rollout |
| Version control | Limited, tool-specific | Git, reviews, migrations, rollback |
| Ownership | Their platform hosts your logic | Repo, database and cloud on your accounts |
The failure mode nobody plans for
Not an outage, a slow drift. Query by query, the tool accumulates rules that exist in no repository, are covered by no test, and are documented nowhere. Everything works, so nobody notices. Then a person leaves, or a rule needs to change under time pressure, and the cost of the last two years of convenience arrives in a single week. The tool did nothing wrong; the logic was simply put somewhere it could not be inspected.
The rule of thumb
Use Retool for screens over systems that already exist and that you already own. Build when the screen is the system, when outside parties need access, when the logic in the tool has become load-bearing, or when the per-user bill is growing faster than the team it serves. As with every tool on this site: the choice belongs to the problem, not to a preference.
Common questions, answered straight
Can we move gradually?
That is the normal path. The usual first move is to pull the logic out from behind the screens into a proper service with tests, and leave the screens where they are. The interface can migrate later, or never, if it is genuinely doing fine.
What about the security side?
Worth checking either way. On one custom platform, a full audit of around 190 API routes turned up 11 access-control gaps, all closed; an external scan of the public surface afterwards found zero high or critical issues. The point is not the numbers, it is that this kind of audit is possible at all when the code is yours to read.
What does the alternative cost?
Free teardown, $490 audit credited toward the build, sprints from $6,900 fixed scope, retainers from $2,900 a month. Published so you can decide before a call rather than during one.
What if the honest answer is to keep Retool?
Then that is what the teardown says. I do not sell Retool work, so the recommendation costs me nothing and is worth something to you.
Related reading
Is your internal tool still a tool?
Free systems teardown: your three biggest time sinks, priced, plus an honest call on which parts should be owned and which are fine as they are. In writing, within 72 hours.
Get your free systems teardown