Show HN: An always-fresh memory that learns your repo, so agents stop re-reading

Category: devtools

Tags: claude-code, mcp-server, agent-memory, codebase-assistant, llm

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

A Claude Code plugin that runs a persistent, cheap large-context model as an MCP server to accumulate codebase knowledge across sessions, eliminating the need for agents to re-read files repeatedly. It innovatively combines passive learning via hooks with a query-answering tool, showing measurable cost reductions for understanding-heavy tasks.

Target audience: backend devs, ai engineers, devops

Repository: https://github.com/shofer-dev/claude-code-live-memory · Python · Apache-2.0 · 1 stars

View on Hacker News