Show HN: Gutsy, subsecond decision model inference on CPU (Jev Style), no API

Category: ai-ml

Tags: local-inference, calibration, decision-making, cpu, llama-cpp, qwen

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

Gutsy is a fine-tuned 0.8B parameter model (based on Qwen) that provides calibrated probability outputs for decision-making tasks, running locally on CPU via llama.cpp. It offers a Jev-style API for yes/no, choice, and score questions, with strong calibration and deterministic results, positioning itself as a private, cost-free alternative to API-based decision models.

Target audience: backend devs, AI engineers, data scientists

Repository: https://github.com/kouhxp/gutsy · HTML · Apache-2.0 · 94 stars

View on Hacker News