Twelve repositories for deploying, counting and checking what you ship

The same job of running your own containers on your own machine, but it is a service installed on the server that you push code to, rather than a tool on the machine you type on; Ubuntu 22.04 and 24.04 or Debian 11 and later only.

The same job of deploying to a server you own, but through a web page with a catalogue of ready-made services instead of a settings file and a command.

Also deploys containers to your own machine and adds a web screen with a web server and free certificates built in, but it runs on Docker Swarm, which is a cluster manager you then have to understand.

The same job of running containers across several machines, but it is real Kubernetes with the central controller and the standard client, which is precisely the weight uncloud avoids.

The same job of placing work on a pool of machines and it scales past ten thousand of them, but it is a server-and-client cluster with its own job format, and its Business Source License 1.1 is not open source.

Also manages applications across several machines you can log in to, but through a web page and an application catalogue rather than a private network and compose files.

The same cookie-free counting with a much better screen, but it needs two separate databases running side by side and the self-hosted edition deliberately leaves out features the paid service has.

The same job and far more widely used, but it does not work at all without a script on the page, and it wants Node.js 18.18 or later with PostgreSQL 12.14 or later.

The same privacy-first counting in a single program and the easiest of the three to install, but it is the free remnant of a product that went paid and its own description promises maintenance and no new features.

Covered in Edition 30. The same small no-personal-data counting in one downloaded program and far more actively maintained, but with no consent banner and no per-visitor encryption; its EUPL-1.2 carries an author's edit to the compatible-licence appendix.

The same audience and the same cookie-free promise, far better maintained, but much heavier to run and the self-hosted edition has fewer features than the paid one.

Also a single small program for privacy-first counting, and feature-frozen by its own maintainers.

The same plain-text double-entry idea and the oldest of the three; written in C++ so it is faster, but it has no Python interface and needs compiling or a system package.

The same model and a nearly compatible file format, and it ships a command line, a terminal screen and a web screen in the box rather than as separate projects, but extending it means Haskell.

Covered in Edition 28. The same double-entry bookkeeping but as a web application with a database and forms, which is easier to use every day and puts the data inside somebody's schema rather than a file you can read.

Covered in Edition 35. The same job with more than 150 indicators and much faster, and the standard choice; the difference is the install, and from 0.6.5 it ships ready-built packages that include the C library.

The same approach of reaching indicators as columns on a table, with more than 50 of them, and actively maintained where ta has gone quiet; needs Python 3.9 or later.

The same set of indicators but recomputed one new price at a time rather than over a whole table, which suits a live feed and is the wrong shape for bulk preparation.

The same allocation job and much easier to pick up, with MIT instead of GPL-3.0, but it solves for one moment in time and has no built-in replay of a policy over history; renamed from robertmartin8/PyPortfolioOpt.

Covered in Edition 33. The same optimisation, also through CVXPY, but presented in the shape of scikit-learn, which makes it familiar if you already work that way; needs Python 3.10 or later.

Covered in Edition 31. It tests thousands of rule combinations very fast, which is a different job: it is a tester, not an optimiser, and its Apache-2.0 plus Commons Clause is not open source.

The same job of checking what goes into and out of a model, at almost the same size and equally active; the difference is that the checks come from a hosted catalogue rather than rules written in a language of its own.

The same safety-filtering goal, but it ships trained models and benchmarks rather than a framework to organise them, and the models are under the Llama Community License, which is not an approved open-source licence.

The same input and output scanning, with 15 checks on the question and 20 on the answer, and archived by its owner on 9 July 2026.

The same job of searching by meaning inside PostgreSQL, and in fact the thing pgvectorscale depends on; the difference is that its indexes are built to live in memory, which is the limit pgvectorscale exists to push past.

Also a disk-friendly similarity index for PostgreSQL by a different method, and the successor to an earlier project by the same group, but dual-licensed AGPL-3.0 and Elastic License v2, so not permissive; renamed from tensorchord/VectorChord.

The same search job done quickly, but as a separate database you run alongside PostgreSQL rather than inside it.

The same job of making a non-text file readable, but as a graphical editor with its own pattern language, and it can write as well as read; GPL-2.0 with its library part under LGPL-2.1.

The same idea of describing a binary format once, but it generates reading code in other programming languages instead of inspecting a file there and then.

Also finds structures inside a binary file, but aimed at pulling files out of device firmware rather than naming every field of a known format.

The same scanning of images and folders and by far the most widely used; the difference is scope, because it also checks configuration mistakes, leaked secrets and cloud settings.

The same static analysis of container images, but built to run as a service that indexes and matches continuously rather than a single command you type.

The same reporting of known holes, but aimed at running Linux and FreeBSD machines and network devices rather than at images, and GPL-3.0 rather than permissive.

The same job across many formats and the library most desktop music players actually use; it is C++ with bindings, so faster and harder to install than a pure-Python package, and dual-licensed LGPL-2.1 and MPL.

The same Python tag editing with a better command-line tool, but it handles essentially only MP3 with ID3, which is the easy case.

The same reading of labels across MP3, FLAC, Ogg, WAV and more, with a far simpler interface and a permissive licence, but it only reads and cannot write a label back; renamed from devsnd/tinytag.