← All repos

claude-mem

Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More

GitHubclaude-mem.ai
aiai-agentsai-memoryanthropicartificial-intelligencechromadbclaudeclaude-agentsclaude-agent-sdkclaude-codeclaude-code-pluginclaude-skillsembeddingslong-term-memorymem0memory-engineopenmemoryragsqlitesupermemory
Browse cluster: Claude AI Agent Frameworks & MCP Tools →
2,385commits
141contributors
5languages

Tech stack & purpose

Claude-mem is a persistent memory system for AI agents that captures tool usage, shell commands, and file edits across sessions, compresses observations using AI, and injects relevant context into future conversations. It functions as a Claude Code plugin and also integrates with Cursor through hook scripts, supporting multiple AI platforms including Claude, Gemini, Copilot, and others. The project is built with Bun as the runtime, uses SQLite for observation storage and Chroma for vector embeddings, and includes a worker service that manages session data and context injection. Claude-mem was created by thedotmack and features integration patterns for both Claude Code's native hook system and Cursor's beforeSubmitPrompt, afterMCPExecution, afterShellExecution, and file-edit hooks.

Community & reference links

Languages

JavaScript
50.2%
TypeScript
47.4%
Shell
1.3%
HTML
1.1%
Dockerfile
0.0%

Contributors (top 30 of 141)