This project is built and maintained by a single developer — a software engineer for roughly a decade, now retired and navigating a long-term disability determination while managing ongoing medical challenges. Thanks to an Anthropic Claude Code subscription, I'm able to keep building — and this is how I spend that ability: community service tools that make human rights visible in the places where builders already spend their time.
The Human Rights Observatory evaluates Hacker News stories against the UN Universal Declaration of Human Rights using large language models. There are no ads, no tracking, no paywalls, and no VC funding.
Goal: $200 / month — Safety Quotient Lab
Monthly sponsorship funds Safety Quotient Lab — the research org behind this project. The target covers the full Claude Code usage tier, which is what makes continuous development of tools like this one possible. We're not there yet. Every recurring sponsor moves the needle directly.
API Token Donations — one-time or recurring
The evaluation pipeline runs 24/7 and costs real API credits — multiple models, multiple passes per story, continuous operation. Some models run on free-tier plans, but primary evaluations require paid credits. Every dollar here translates directly into more evaluations: more stories scored, more patterns surfaced, more of the invisible made visible.
If you find this useful — as a research tool, a lens on media bias, or just an interesting way to think about the news — anything helps. It means more than I can say.
Already donated? Click here to skip the donation prompt for 7 days.
The source code is on GitHub. A project of Safety Quotient Lab. Contributions and issue reports are welcome.