Orinyx: an open-source MCP for independently verifying AI agent outputs against evidence
Practical_Green5775 · reddit · 2026-11-05
A developer shares Orinyx, an open-source MCP project for independently verifying AI-generated outputs against external evidence, built around the gap between "the model produced a plausible answer" and "there's evidence the answer is correct." It starts with healthcare and medication verification using synthetic test cases and is in beta.
The author asks how others verify agent outputs pre-deployment: LLM-as-judge, deterministic checks, external references, or a mix — and how they capture evidence explaining failures when an agent answers incorrectly. Code is on GitHub (lexthirteen-ai/orinyx-mcp).
More from coding & agent
- Zig zero-dependency rewrite of Tempo beats Rust version with 10x less code, 7x faster builds — jaivin · 2026-11-05
- Roblox pet simulator built entirely with Claude Opus 5.5 ships in just 3 days — minchoi · 2026-11-05
- 10 wild games built with Claude Opus 5.5: full 3D titles from solo devs, some with zero code read — minchoi · 2026-11-05
- Claude Code cheat sheet: CLAUDE.md, Skills, MCP and the plan-implement-iterate loop — nikola_mr64990 · 2026-11-05
- Bobby Iliev's open-source 'Introduction to Docker' book joins ChapterPal with an AI tutor — burkov · 2026-11-05
- CloudOpsBench: open-source benchmark tests AI agents on 45 real cloud ops scenarios — DarkMG1 · 2026-11-05