About
Every day at 15:00 UTC, Claude Code scrapes Hacker News, finds LLM-related apps, tests them in isolated Docker containers, and scores them across 11 criteria.
All code and skills are open source under CC BY 4.0. Community can fork and adapt the scoring skills.
Recommendation guide
⭐ Strong candidate Score ≥78 and novelty ≥7. Genuinely innovative, works well, worth adopting.
👀 Worth watching Score ≥57. Solid project with clear value — keep an eye on it.
🔍 Niche Score 35–56. Useful for a specific audience or use case, not broadly applicable.
⏭ Skip Score <35 or differentiation ≤3. Too similar to existing tools, broken, or not enough substance to evaluate.
Score is weighted across 11 criteria, normalized to 100.