1
0
Fork
You've already forked open-slopware
0
[note: Just forking this in case it goes away again] Free/Open Source Software tainted by LLM developers/developed by genAI boosters, along with alternatives.
2026年01月10日 05:42:28 +04:00
evidence add zbus 2026年01月09日 10:58:30 +01:00
img upload new jj info 2026年01月09日 01:14:35 +01:00
LICENSE.txt Add CC-BY-SA-4.0 license 2026年01月10日 05:42:28 +04:00
README.md add alternatives, yaayyyy 2026年01月10日 05:42:28 +04:00

Open Slopware

Free/Open Source Software tainted by LLM developers/developed by genAI boosters, along with alternatives.

The intention of this list is to raise awareness of this technology's usage in popular software, as well as give people informed alternatives that they can reach for when they want to make decisions for themselves. This is not a list created so you can go and give these projects trouble for their decision. If you want to file a complaint about it with them, we consider that acceptable but ask that it be done respectfully and constructively.

For more information on why we think this list matters, please see Why not LLMs?.

Contributing

To contribute to this list, file a pull request adding it to the appropriate category, along with any known alternatives. The pull request must include links to evidence of such (social media posts, screenshots, commit history, links to CLAUDE.md, etc). Title the pull request "Add <software-name>" or such.

Criteria

The criteria for being included in this repository are any of:

  1. One or more core/prominent developers openly use LLMs in any level of supportive capacity to develop this software specifically. Accepting drive-by or minor contributions that happen to use LLMs is NOT qualifying. This may change in the future.
  2. The software itself integrates genAI features of any kind, internally or externally.

Removal

Any project here may be removed as long as the following become true, even if prior LLM-authored code remains:

  1. The project establishes a "No genAI" policy, where the core developers commit to no longer using LLMs for developing that project in particular and no longer accepting obvious LLM-authored contributions, and state the policy prominently in their repo.
  2. Any genAI functionality of the project itself is removed in its entirety. "Kill switches" are insufficient.

Tainted Software

Browsers

Name Evidence Notes
Firefox Blog post collecting details
Floorp (Firefox fork) GitHub discussions - LLM folders in structure, LLM "contributors", own MCP server, "there is always a possibility that we will implement our own AI features"
WaterFox (Firefox fork) @cf's toot There's a sort of negative blogpost on LLMs in browsers but otherwise the project seems to welcome it on the whole.

Alternative browsers

  • LibreWolf (Firefox fork) has made statements and efforts to remove Mozilla's push for "AI". They also acknowledge that the notion of "apolitical" technology is a futile one.
    • For Android, while not officially related, some of the same developers make IronFox
  • GNOME's Epiphany (WebKit based) does not currently have a stance, but GNOME developers on the whole show aware anti-"AI" sentiments and roll out wide reaching rules against it.
  • Servo (New engine, formerly owned by Mozilla, now by a co-op) is not yet daily driver ready, but has strong protections for when it is some time in the future.

Command-Line Shells

Name Alternative(s) Evidence Notes
nushell elvish shell, fish shell Integrated MCP server, core developer sets Claude as a co-author
yash dash shell, but there are no other major shells fully targeting POSIX Assigns and auto-reviews MRs with LLMs

Chat

Name Alternative(s) Evidence Notes
Zulip Delta Chat Integrated an LLM "topic summary" feature
Mattermost Delta Chat Distributes an "agents" plugin, with several LLM features, core developers commit LLM-generated code

Computer Graphics

Photography

Name Alternative(s) Evidence
RapidRAW From the README: "with the support from Google Gemini." Roadmap includes adding GenAI.
RawTherapee Stick to version 5.12 Use Copilot for review PR7518, PR 7435, PR 7378 and code changes commit c4b0400

Terminal Emulators

Name Alternative(s) Evidence
Warp Advertised as an "agentic development environment."
Kitty Alacritty, WezTerm Copilot co-authored commits. Author also responsible for adding LLMs to Calibre
iTerm2 Deep integration of OpenAI-backed features. Killswitch only.
Ghostty Blog post "Vibing a Non-Trivial Ghostty Feature", Post on Mastodon 1, Post on Mastodon 2, Post on Mastodon 3, Post on Mastodon 4, Post on Mastodon 5

Developer Tools

Name Alternative(s) Evidence
cargo-nextest Just use cargo test CLAUDE.md and AGENTS.md in repository
jujutsu continue using git itself A core dev occasionally uses Claude to generate local test cases., (Discord link)
just Lists an MCP server in its User Manual.
oh-my-posh Starship AGENTS.md file in main repository
oh-my-bash Starship All PRs are reviewed with "Qodo" Example PR
ohmyzsh Starship Added Claude and Grok as search engines
Taskfile Uses Copilot for PRs

Programming Languages

Name Alternative(s) Evidence Notes
.NET See below .NET Runtime PRs, .NET SDK PRs
Julia Python (w/ appropriate scientific libraries) AGENTS.md in main repository, PRs "written by Claude"

Note on alternative languages

Suggestions for alternative programming languages are necessarily more complex than many other categories considered here, as that's one of the core technical choices to be made when starting a new software development project.

Depending on your needs and technical concerns, you may find any number of different alternative languages useful. For example, when considering alternatives to .NET, Java or Scala may be interesting candidates from the perspective of being similar VM-based frameworks, while Python or Rust may be interesting alternatives in other ways.

Editors

Name Alternative(s) Evidence Notes
Zed Lapce, other popular editors like neovim, emacs, etc They prominently advertise genAI features as a core part of their product. See their website. There's a killswitch to disable genAI features entirely, but it's likely the code itself is still heavily LLM-authored.
VS Code VS Codium* removes the AI features from VS Code It's literally their main advertising tagline, and they publicly use LLMs in their public repository. Most AI "features" are enabled by default, and some cannot be disabled. *VS Codium is still written with LLMs since Microsoft owns it

Accessibility Tools

Name Alternative(s) Evidence Notes
espeak-ng flite Dev uses copilot for review and is setting up copilot instructions espeak still exists, but is heavily outdated and very bad from a modern standpoint, so it cannot be considered as an alternative. flite users may also wish to install speech-dispatcher-flite.

Microcontroller Tools and Libraries

Name Alternative(s) Evidence Notes
FastLED .claude and .cursor folders in main repository

Service Monitoring

Name Alternative(s) Evidence
Uptime Kuma Dev uses copilot for review

Libraries

Rust

Name Alternative(s) Evidence Notes
facet Use more specific deserialization libraries and avoid macros that slow things down anyway. serde itself is tainted by its Palantir-defending author. nanoserde could be suitable for certain use cases. Core developer openly uses LLMs for library dev
ratatui iocraft and cursive. console and its family of libraries may also be enough for some use cases. Core developer has stated their use of AI for rust code
yaxpeax Consider linking to binutils' libopcodes. Core developer learned something from Claude once
zbus CLAUDE.md, core developer talking about their AI policy (screenshot)

CSS

Name Alternative(s) Evidence Notes
tailwindcss twind Sponsorship page states explicitly that the founder uses AI to write code and even sells access to his "Cursor/Claude/AGENTS.md rules"
daisyui Homepage advertises MCP server and disyplays llm.txt in docs

JavaScript / TypeScript

Name Alternative(s) Evidence Notes
Zod TypeBox, ArkType (both in grey area) CLAUDE.md, AGENTS.md, .cursorrules in GitHub repo Is also sponsored by AI startups
Valibot TypeBox, ArkType (both in grey area) Uses LLM's in GitHub issues and PR's; prompts folder; LLMs.txt

Game Engines

Name Alternative(s) Evidence
UPBGE 0.5 UPBGE 0.36.1 seems to be slop-free 0.5 release notes state the use of AI in plain language

Mathematics software

Name Alternative(s) Evidence
Golly 5.0 release notes say the app icon is generated by Claude.

Services/Utilities

Name Alternative(s) Evidence
systemd sysvinit, OpenRC, GNU Shepherd, among others. Systemd maintainer Daan De Meyer vibe codes contributions and requests reviews from Copilot. Also, the repository has a copilot-instructions.md
AliasVault Main instance uses "AI" "generated" avatar
Bluesky Mastodon CTO discusses using Claude in development
Forgejo (and by extension Codeberg) Git patches via E-Mail Core contributor w/ 34+ merged PRs admits to using LLMs in Forgejo contributions without disclosure
Calibre 8.16.2 BookLore
Clbre, a Calibre fork without AI
Arcalibre, a Calibre hard-fork from before AI features
Added "Discuss selected book(s) with AI" feature
Bitwarden Include CLAUDE.md
Calibre 8.16.2 BookLore
Clbre, a Calibre fork without AI
Arcalibre, a Calibre hard-fork from before AI features
Added "Discuss selected book(s) with AI" feature
Ghost WriteFreely Include configurations for Claude and cursor; images on website seem generated
Home Assistant Include configurations for Claude and Copilot; Use Copilot for review and code changes
KeepassXC Use official KeePass 2. It is a .NET application, but Mono or Wine can be used for Linux, MacOS, BSD, ...
KeePassXC 2.7.9 was released before the statement, and was awarded 3 year security Visa by the French National Cybersecurity Agency (ANSSI) for a First-level Security Certification (CSPN), valid in France and Germany, under report No. ANSSI-CSPN-2025/16 (archive.org, certificate).
Explicitly stated in official blog post.
Use Copilot to generate PRs: Example 1, Example 2, Example 3.
MinIO Garage Advertises AI as their primary focus on their website.
Nextcloud Nextcloud Assistant; Nextcloud Desktop has AGENTS.md
PeerTube Includes copilot instructions
Wagtail CMS WriteFreely Include AGENTS.md; Wagtail AI
WordPress WriteFreely Include AGENTS.md; has Copilot PRs
Rspamd .claude/, .cursor/

Operating Systems

Name Alternative(s) Evidence
Linux Kernel BSD? Honestly this one's tough Maintainers are explicitly trying out the technology and considering LLM-based reviews

Why not LLMs?

Sketching things in; citations needed.

Confabulation

LLMs work by generating something that looks like what the prompt has asked for. They're prediction engines; after generating the first best-match token, they search for the next token that's most likely to come next.

Sometimes, this does what you want. Frequently, it's something wrong. LLMs have no "understanding" or "reasoning" despite the marketing, and they have no way of knowing what's true or not. They don't "know" anything, they're just generating something like an answer.

TODO: links/discussion of the guy who had an LLM delete everything and then whoopsie about it, article about an LLM "reviewing" someone's documents but not actually every reading them, etc.

Labor

Because LLMs generating things that look like answers, executives who work by accepting things that look like answers think they're magic.

  • people are losing their jobs because executives and managers are investing resources in LLMs instead of salaries; they believe they'll be able to regain the productivity losses by using LLMs and cheaper employees
  • it's rare for people to actually lose their jobs directly because of LLMs, but it does happen:
    • Tailwind's business is driven by people accessing their documentation; their business has tanked because LLMs just answer people's questions (or do they?), so nobody goes to their website and buys. They laid off 75% of their staff in 2026-01.

This genius idea of using fresh out of college (or whatever) junior employees to do everything with the help of LLMs will result in a complete lack of experienced employees. Nobody using LLMs is learning anything other than how to make them do what you want.

Who's going to fix the security problems introduced by code nobody on the team understands? Who's going to be able to to fact check when "all" information is presented from LLM answers that may have been tweaked by the owners?

Environment

LLMs are insanely inefficient, requiring enormous amounts of datacenter resources including:

  • electricity
  • water
  • specialized GPUs
  • RAM
  • storage
  • they need to constantly re-train on everything

Theft

LLMs are trained on the contents of the Internet, including pirated content. This is bad because:

  • it ignores copyright and licensing completely
  • LLM scrapers ignore robots.txt and other mechanisms intended to prevent bots from overwhelming a site's resources
  • it includes horrendous content like revenge porn and child sexual assault material (CSAM)

License

This repository is licensed under the Creative Commons Attribution Share Alike 4.0 International license. Please see LICENSE.txt for more information.