What is Memvault?
Memvault is a local-first conversation archive and memory system for AI agents.
The Problem
AI conversations are valuable assets:
- They contain your reasoning process, decisions, and context
- Agents lose memory between sessions
- Platform-hosted conversations are not portable
- No way to search across all your AI interactions
The Solution
Memvault gives you complete ownership of your conversation data:
- Archive Everything: Automatically capture conversations from Claude Code, Codex, and other agents
- Search Locally: Full-text search across all archived conversations
- Agent Memory: Let agents access past context through MCP integration
- Portable: All data stored locally in open formats
- Private: No cloud dependency, no data leaves your machine
Key Features
Local-First Architecture
All data stays on your machine. Sync across devices using your own cloud storage (iCloud, Dropbox, etc.) — Memvault never sees your data.
Multi-Agent Support
Works with:
- Claude Code (terminal and desktop)
- Codex by Anthropic
- Future: OpenCode, custom agents
Memory System
Agents can:
- Read past conversations
- Build long-term context
- Reference previous decisions
- Learn from history
Full-Text Search
Instant search across:
- All conversations
- All agents
- Markdown memories
- Chinese and English content
Use Cases
- Development Context: Keep track of architecture decisions and code discussions
- Research: Build a searchable knowledge base from AI conversations
- Documentation: Extract insights from past conversations
- Continuity: Provide agents with long-term memory across sessions
Next Steps
- Getting Started - Install and configure
- Core Concepts - Understand the architecture
- CLI Reference - Command-line usage