LogixLoops
We standardised on a small set of tools and stopped relitigating it every project. React, Next.js and TypeScript on the front. Node.js or Go behind them. Postgres, Kubernetes, Terraform. All open source, all boring, all still here in five years.
Twenty years of tooling, in current use
Every tool here is open source or an industry-standard managed service. There is no LogixLoops runtime, no proprietary framework, and nothing you would have to keep paying us for. If you replace us tomorrow, the next team recognises everything in the repository.
Anyone will list the tools they like. The more useful list is the one they avoid, and why, because that is the list that tells you what your codebase will look like in three years.
A distributed system is a debugging tax you pay forever. We start with a well-bounded modular monolith and split the parts that genuinely need to scale or deploy independently. Most products never need the second step.
If a library will not plausibly be maintained for as long as you intend to run the product, it is a migration you have scheduled without noticing. Boring, widely-adopted tools cost less over five years than exciting ones cost in year two.
Nothing we build depends on a LogixLoops package, a private registry, or a platform you cannot leave. The whole stack is open source or a standard managed service, so replacing us is a staffing decision rather than a rewrite.
A great deal of what gets sold as AI is a SQL query, a rules engine, or a well-indexed search. We reach for a model when the problem is genuinely probabilistic, and when it is, we say what it costs per call before you commit.
Three things that run automatically, on every change, whether or not anyone is watching.
Typed end to end in TypeScript or Go, linted on commit, and held to a coverage threshold the pipeline enforces. A change that drops below it does not merge, so nobody has to remember to care.
Containerised with Docker, described in Terraform, orchestrated by Kubernetes where the load justifies it. The environment is a file in the repository, not knowledge in somebody's head.
Tracing, metrics and structured logs ship with the first release rather than after the first outage. Alerts fire on user-visible harm, not on a CPU graph nobody agreed on.
Send us the repository and the pain. You get a written read on what is worth keeping, what is quietly costing you, and what order to fix it in.