Show HN: Lore – LLM proxy for coding agent context and memory management

Category: ai-ml

Tags: llm-proxy, ai-agents, memory-management, context-engineering, local-first, vector-search

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

Lore is a local-first LLM proxy that manages context and memory for coding agents like Claude Code and Codex. It extracts decisions, gotchas, and patterns from sessions, curates them into a git-portable .lore.md file, and injects relevant knowledge back into future sessions, aiming to reduce context loss and token costs.

Target audience: AI engineers, backend developers, devops using coding agents

Repository: https://withlore.ai/

View on Hacker News