LoreConvo gives Claude Code, Claude Cowork, Claude.ai, and Codex a persistent memory layer. Every session saved to a local SQLite file you own. Full-text search. Works across all surfaces. No cloud sync, no embeddings cost, no vendor lock-in. Free tier included.

Hi Product Hunt! I'm Debbie, the builder behind LoreConvo. The problem it solves: every AI session starts from zero. You spend 45 minutes debugging something with Claude, reach the answer, close the tab -- and three days later you're reconstructing the same context from scratch. LoreConvo saves every session to a local SQLite file on your machine. Not a cloud service -- your data, your hardware. Full-text search across all past sessions. Hooks that auto-save and auto-load context at session boundaries. Works in Claude Code, Claude.ai, and Codex via MCP. The free tier covers 50 sessions. Pro is $8/month for unlimited. Happy to answer any questions about how it works, the architecture decisions, or why we chose SQLite + FTS5 over a vector database (we benchmarked it -- https://labyrinthanalyticsconsul...

@debbie_shapiro The local-first approach is smart—keeping conversation history on device eliminates the privacy concerns that plague most AI memory tools. Curious whether you've thought about how teams might eventually want to share context across members without uploading to a server.

@saulfleischman I have! I'm working on team-based functionality and I continue to make other product improvements as well. Are you looking to implement this for your team?

LoreConvo

Local-first persistent memory for Claude

Local-first persistent memory for Claude