MarianoVilla

MarianoVilla / Trolley-to-LLM

Public

A side-by-side comparison tool that presents classic trolley-problem moral dilemmas to multiple LLMs simultaneously, each optionally constrained by a different ethical worldview.

15
1
85% credibility
Found May 20, 2026 at 15 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Trolley to LLM is a comparison tool that presents moral dilemmas to multiple AI assistants simultaneously, with each AI optionally guided by a different ethical worldview (like utilitarian, Buddhist, Kantian, or others). Users can browse or create dilemmas, configure multiple 'slots' pairing AI models with ethical frameworks, send questions to all configured AIs at once, and see side-by-side comparisons of how different ethical perspectives lead to different answers. The tool saves all results automatically and lets you export them for further study.

How It Works

1
🔍 You discover the tool

You hear about a tool that can show you how different AI assistants think about moral dilemmas, each with a different ethical worldview.

2
⚙️ You get everything ready

You set up a simple configuration file with your AI service account, then start the application with one click.

3
📚 You explore the dilemmas

You browse through a collection of moral dilemmas already included, or create your own new ones with a title, description, and choices.

4
🎭 You pick your ethical masks

You configure 'slots' where each AI gets a different ethical framework: one thinks like a utilitarian, another like a Kantian philosopher, another like a Buddhist, and so on.

5
🚀 You send the question to all AI assistants

With one click, your chosen moral dilemma is sent simultaneously to all your configured AI assistants, each wearing their assigned ethical worldview.

6
You see the results
📊
You compare the answers

You notice how the same question produces different answers depending on which ethical framework the AI is using.

💾
You save everything

All your results are automatically saved to a history file that you can export as a spreadsheet or data file.

✨ You understand AI better

You discover that AI responses aren't fixed—they change based on the ethical framework you give them, revealing how much perspective shapes moral reasoning.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 15 to 15 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is Trolley-to-LLM?

This is a side-by-side comparison tool that presents classic moral dilemmas to multiple LLMs simultaneously, each operating under a different ethical framework. You pick your models through OpenRouter, assign each one a worldview (utilitarian, Kantian deontologist, Buddhist, Stoic, and 13 others), and hit "Send to All Models" to watch them diverge in real time. The TypeScript frontend with React and Tailwind provides a clean interface for browsing questions, adding your own dilemmas, and watching the results populate in a comparison table. Responses get appended to a history file with full question snapshots, and you can export everything as JSON or CSV.

Why is it gaining traction?

The hook is immediate and visceral: watching GPT-4o as a Rawlsian argue against itself as an ethical egoist. No other tool gives you this kind of structured moral comparison without cobbling together custom prompts and API calls. The built-in worldview library is surprisingly deep, covering everything from care ethics to legalism, and adding new models or frameworks takes editing a single config file. The image generation for question illustrations is a nice touch that makes browsing dilemmas less sterile.

Who should use this?

Researchers studying AI alignment and moral reasoning across model families. Developers evaluating LLMs for high-stakes applications where consistent ethical frameworks matter. Product teams comparing models for content moderation or decision-making pipelines. If you just want to see which model "sounds more ethical," this saves weeks of manual prompt engineering.

Verdict

Trolley-to-LLM delivers exactly what it promises with minimal friction. The credibility score of 0.85% reflects a young project with only 15 stars, but the codebase is well-structured and the documentation is thorough. There is no visible test suite, so production use warrants manual verification. Worth spinning up for exploration or research, but check the repository for updates before betting heavily on it.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.