Show HN: Recursive Human Intelligence: Actionable feedback from your LLM chats

Category: devtools

Tags: llm, productivity, chat-analysis, privacy, java, desktop-app

Score: 5.8/10 (Innovation: 5, Technical: 6, Documentation: 7, Utility: 5)

RHI is a Java desktop application that analyzes your chat transcripts from AI coding assistants like Claude Code and Codex to surface recurring behavioral patterns and actionable self-improvement feedback. It uses an LLM to group your requests and generate concrete suggestions, with a focus on privacy controls and local-first processing. The project is interesting for its unusual angle of applying LLM analysis to the user's own prompting habits rather than to code.

Target audience: developers who use AI coding assistants daily and want meta-feedback on their own habits

Repository: https://github.com/andrearitossa/recursive-human-intelligence · Java · Apache-2.0

View on Hacker News