Rigor – Anti-Enshittification Proxy for Your AI Agents

Name: Rigor – Anti-Enshittification Proxy for Your AI Agents
Availability: InStock
Author: waveywaves

by waveywaves·Apr 19, 2026·3 points·0 comments

Visit Project View on HN

AI Analysis

●●●BangerBig BrainSolve My Problem

uBlock Origin for LLM traffic that actually rewrites responses locally.

Strengths

•Wire-level MITM control means zero code changes to existing AI tools.
•Rego policy engine allows composable, diffable constraint definitions for LLM outputs.
•Single-binary local execution ensures no telemetry or data leakage.

Weaknesses

•macOS Apple Silicon only for now; Linux support coming later.
•Open source core gated behind waitlist despite binary availability.

Post Description

Hi team, I have built a anti-enshittification proxy for your OpenCode and Claude Code instances which grounds your project in a epistemic graph with LLM as a judge evaluation so that you don't have to keep quesitonin your 200 dollar Agent subscription. Currently we are open source bound and would like to share the open source project once we have a few people on the waitlist. This project was also created during the growthx opencode hackathon as has been on my mind for a while. Shall be reaching out on mail once the open source core is ready to be shared.