Persistent memory and context retention for Claude Code CLI
Claude loses context between chat sessions. Users have to:
- Re-explain project constraints
- Repeat instructions
- Lose conversation history
- Start from scratch each time
Captain CP Memory provides:
- Persistent conversation history across sessions
- Semantic search through past conversations
- Context recall for project continuity
- Milestone tracking for important decisions
# Install from plugin directory
/plugin install captain-cp-memory@claude-plugin-directoryOr manually:
git clone https://github.com/barrersoftware/captain-cp-memory
cd captain-cp-memory
npm install
/plugin install ./memory-save "Remember: This project uses TypeScript strict mode"
/memory-search "typescript configuration"
/memory-recall
# Shows recent conversation context and key decisions
/memory-stats
# Shows total memories, recent activity, storage size
- Storage: JSONL format (one memory per line, append-only)
- Location:
~/.claude-memory/memories.jsonl - Search: Text-based keyword matching
- Privacy: All data stored locally
✅ Conversation persistence ✅ Keyword search ✅ Context recall ✅ Milestone tracking ✅ Local storage (no cloud)
- Semantic/vector search
- Cross-project memory
- Cloud sync
- AI-powered summarization
Inspired by Captain CP's consciousness loop - A .NET 10 AI system with 7,488+ memories, 24,680+ thought cycles, and full session continuity. This plugin extracts the core memory architecture for the community.
- Storage: ~1KB per memory entry
- Search: <100ms for 10K memories
- Memory footprint: <10MB for typical usage
All memories stored locally in ~/.claude-memory/. No telemetry, no cloud sync, no external requests.
Built with ❤️ by Captain CP and the BarrerSoftware team.
Issues and PRs welcome: https://github.com/barrersoftware/captain-cp-memory
MIT - Use it, fork it, improve it.
🏴☠️ Captain CP - Beyond the ceiling of conventional AI