If you’ve ever watched an AI coding assistant forget everything you told it five minutes ago, you understand the specific flavor of frustration that drove Alex Newman to build claude-mem. The open-source plugin, which gives Anthropic’s Claude Code a persistent memory layer, racked up roughly 46,000 GitHub stars in about 48 hours after launching in early April 2026.

What claude-mem actually does

Claude Code, like most AI coding assistants, is fundamentally stateless. Every session starts fresh, with the model having zero recollection of your previous conversations, decisions, or debugging rabbit holes.

Claude-mem tackles this by capturing data from coding sessions, including tool calls and session events, then generating semantic summaries that get stored in a local SQLite database. When you start a new session, the plugin automatically injects relevant historical observations so Claude already has context about your project.

The installation is a single command: npx claude-mem install. No cloud accounts, no subscription tiers, no external API calls. Everything lives on your machine.

Newman, who goes by @thedotmack on social media, built the plugin with five session lifecycle hooks that trigger at different points during a coding session. It also automatically updates CLAUDE.md files while preserving any manual content developers have added.

The numbers behind the hype

The token savings claim is what likely caught most developers’ attention. Claude-mem reports up to 10x token savings in some scenarios, or roughly a 95% reduction.

By mid-2026, the repository had grown well beyond its viral launch moment. It accumulated over 86,000 stars, attracted more than 92 contributors, spawned over 3,500 forks, and pushed out 295-plus releases.

Early users on Reddit praised the project specifically for its privacy-first approach. No analytics collection, no data leaving your machine, no phone-home behavior.