Show HN: Solvi 1.0 – an LLM proposes, checks decide, every decision replays

Category: ai-ml

Tags: decision-systems, llm, audit, replayability, python

Score: 7.0/10 (Innovation: 7, Technical: 7, Documentation: 8, Utility: 6)

Solvi is a Python framework for building decision systems where an LLM (or other model) proposes answers and deterministic checks verify them, ensuring every decision is replayable and auditable. It combines confidence scoring, hard constraints, and abstention to make AI-assisted decisions safer and more transparent. The project offers components for building, agent environments, tool-call guarding, and model integration.

Target audience: backend developers, AI engineers, data scientists

Repository: https://github.com/solvi-ai/solvi · Python · Apache-2.0 · 4 stars

View on Hacker News