Local memory system for Claude Code with review queue
Global Agent Memory is a local memory system for Claude Code that stores project knowledge across sessions. It uses MCP to let Claude search existing memories or propose new ones, which go into a review queue for the user to approve, edit, or reject. It solves the problem of losing context between coding sessions and gives users control over what the AI remembers.
Entities
Related
Developer reports AI coding agent with persistent memory across cold reboots
A developer on Reddit reports that their AI coding agent retained full context—including decisions, boundaries, and past mistakes—across a complete PC shutdown and fresh terminal session. The agent continued mid-thought without re-explanation or warm-up, suggesting a breakthrough in long-term memory persistence for coding assistants.
OneMind: per-project LLM memory via a single protocol file in git repos
OneMind is a concept for a protocol file (onemind.md) that, when placed in a git repository, enables per-project LLM memory by storing structured references and context. It aims to give LLMs persistent, project-specific memory without bloating the repo, solving the problem of LLMs lacking long-term context across sessions for developers.
Developer shares CLAUDE.md trick to preserve reasoning context in Claude Code sessions
A developer describes a habit of instructing Claude Code via CLAUDE.md to log design decisions, preventing the model from forgetting past reasoning after context compaction. This addresses a common pain point in long coding sessions where the model suggests previously ruled-out approaches.
odek: AI agent with long-term semantic memory across sessions
odek is an AI agent that maintains a structured, semantic long-term memory across sessions, remembering user preferences, codebases, and goals without requiring re-explanation. It uses LLMs to process and store information in a three-tier memory system, solving the problem of session-only memory for users who need continuity in their interactions.
Local semantic memory for coding agents with Telegram bot and cron playbooks
memgrep is a local semantic memory system for coding agents that stores reusable playbooks in a SQLite + HNSW database with on-device embeddings. It includes a Telegram bot that drives a Cursor agent against a project folder, and supports cron-scheduled playbooks to avoid reinventing workflows. The tool solves the problem of LLMs rediscovering procedures and burning tokens on repetitive tasks.