Hacker GitHub

← All repos

anulum/director-ai

View on GitHub ↗ ★ 2 Python

Real-time LLM hallucination guardrail — NLI + RAG fact-checking with opt-in claim-level streaming contradiction halt. Drop-in for any LLM backend.

4 mentions on Hacker News by 1 person: 1 story, 3 comments · first 2026-02-26 · latest 2026-02-27

Mentions per year 4 mentions · busiest year 2026 (4)
20082026
AllStoriesComments NewestTopOldest
  1. …interceptors, score caching, 8-bit NLI, bge-large, LangGraph/Haystack/CrewAI support Repo: https://github.com/anulum/director-ai Changelog: https://github.com/anulum/director-ai/releases/tag/v1.2.0 Would love feedback on the new bits —…
    comment by anulum · 2026-02-27 · in Show HN: Director-AI – token-level NLI+RAG
  2. …— feel free to fork and run your own tests. Would love to hear what you find. Benchmarks section: https://github.com/anulum/director-ai#benchmarks
    comment by anulum · 2026-02-26 · in Show HN: Director-AI – token-level NLI+RAG
  3. …you tried frontier self-critique in real streaming agents? What broke for you? Repo benchmarks: https://github.com/anulum/director-ai#benchmarks
    comment by anulum · 2026-02-26 · in Show HN: Director-AI – token-level NLI+RAG