LinkHarvestDigest

Show HN: Veru—The Open-Source AI Citation Auditor Built on OpenAlex Artificial intelligence tools have become powerful research companions, but they come with one glaring weakness: citations. Large language models often hallucinate references , fabricate sources, or misattribute academic work. While AI writing tools evolve rapidly, reliable citation verification still lags behind. That gap is exactly what Veru , an open-source AI citation auditor , was built to solve. Veru connects AI-generated references to OpenAlex , the open, community-driven database of academic papers, authors, institutions, and journals. Instead of hoping your citations are real, Veru checks them automatically — flagging missing papers, mismatched metadata, incorrect DOIs , inconsistent titles, and fabricated sources. The result is a reliable auditing pipeline for researchers, students, technical writers, and anyone generating citations through AI. This article explores Veru’s purpose, how it works, what probl...

     Alignmenter: A Breakthrough Framework for Measuring Brand Voice Consistency in AI Systems If you're building conversational AI systems—especially AI copilots—you know how critical it is to maintain a consistent brand voice across model versions. But how do you quantify something as subjective as “sounds right”? Enter Alignmenter , a robust open-source framework designed to make persona alignment measurable, reproducible, and CI/CD-friendly.  What Is an Alignmenter? Alignmenter is a Python-based toolkit that evaluates conversational AI responses across three core dimensions: Dimension Description Authenticity Measures style similarity (via embeddings), trait patterns (logistic regression), lexicon compliance, and optionally uses an LLM judge Safety Uses keyword rules, a distilRoBERTa classifier, and optional LLM judge to flag unsafe content Stability Calculates cosine variance across response distributions to assess consistency This multi-pronged approach ensure...