Show HN: LLM memory without context bleed; 100% precision vs. <10% vector search

Category: infrastructure

Tags: llm-memory, local-proxy, ai-context

Score: 7.3/10 (Innovation: 7, Technical: 8, Documentation: 7, Utility: 7)

Tenure is a local proxy that provides persistent AI memory across tools and sessions by extracting structured 'beliefs' from conversations and injecting relevant context into LLM requests. It stands out by using BM25 retrieval with scoped hard boundaries instead of vector search, achieving clean separation between projects and eliminating context bleed.

Target audience: backend devs, devops, data engineers

Repository: https://tenureai.dev/ · TypeScript · MIT · 10 stars

View on Hacker News