An open-source MCP server that connects Claude Desktop, Cursor, ChatGPT, and 113+ AI clients to your entire conversation history. One install. Works everywhere.
PyPI package coming soon — pip install aimemory-mcp-server will work once published
One command. No configuration files. No accounts.
✓ Python 3.8+ ✓ Works on macOS, Linux, Windows ✓ No API keys needed
📦 PyPI package coming soon — use GitHub install for now
Everything your AI assistant needs to remember, search, and manage your conversations.
Full-text search across all your saved conversations with FTS5 syntax support. Filter by platform, date, or tags.
Save new conversations, insights, or memory snippets with automatic tagging and source attribution.
Browse your memory library with tag filtering, date ranges, and pagination. See what your AI remembers.
Retrieve a specific memory by ID. Perfect for referencing exact conversations or insights your AI saved earlier.
Edit existing memories — update content, add tags, correct details. Keep your knowledge base accurate.
Remove outdated or irrelevant memories permanently. Full control over your data.
Get total memory count, recent activity, and tag distribution. Monitor what your AI remembers at a glance.
Backup all memories to JSON. Export your entire knowledge base for migration or safekeeping.
Import memories from JSON backup. Batch-load conversations with automatic duplicate detection.
Save multiple memories at once. Perfect for extracting key takeaways from a conversation in one call.
List all unique tags with usage counts. Discover what categories of memories you have stored.
Delete all memories. ⚠️ Use with caution — export first if you want to keep a backup.
One server, every client. Same config everywhere.
Add to claude_desktop_config.json:
{
"mcpServers": {
"ai-memory": {
"command": "aimemory-mcp-server"
}
}
}Restart Claude Desktop. Ask: "Search my memory for React performance tips"
Go to Settings → MCP → Add New MCP Server:
AI Memorystdioaimemory-mcp-server{
"mcpServers": {
"ai-memory": {
"command": "aimemory-mcp-server"
}
}
}In your MCP settings, add:
{
"mcpServers": {
"ai-memory": {
"command": "aimemory-mcp-server"
}
}
}Also works with: Zed, Continue, Augment, Aider, and 100+ more MCP clients
View full documentation →Choose local or cloud — both give your AI persistent memory.
Runs entirely on your machine. Your data never leaves your computer.
pip install aimemory-mcp-server
Use our hosted endpoint. Upload conversations via the web UI, then search from any MCP client.
Endpoint: aimemory.pro/api/mcp
The Model Context Protocol is the USB-C of AI — one standard, infinite connections.
One server works with Claude, Cursor, ChatGPT, and 113+ clients. No custom integrations per platform.
Your AI remembers conversations from ChatGPT, Claude, DeepSeek, and Gemini — all searchable from one place.
Cross-platform memory format is an industry gap. AI Memory is building the "SMTP of AI memory" — and you can use it today.
The only MCP server with cross-platform conversation search.
| Feature | AI Memory MCP | Mem0 | Custom MCP |
|---|---|---|---|
| Setup Time | 10 seconds | 30+ minutes | Hours |
| Cross-Platform Search | ✅ 4 platforms | ❌ Single source | ⚠️ Manual |
| MCP Standard | ✅ Native | ❌ REST API only | ✅ DIY |
| Offline Support | ✅ Full | ❌ Cloud only | ⚠️ Depends |
| Pricing | Free | Free tier + $24M VC | Dev time |
| Consumer-Friendly | ✅ Zero config | ❌ Developer-only | ❌ Code required |
An MCP (Model Context Protocol) server for AI memory is a tool that gives AI assistants like Claude Desktop and Cursor persistent memory by connecting them to your conversation history. AI Memory's MCP server lets any MCP-compatible AI search, save, and retrieve memories across all your AI conversations.
Install with one command: pip install aimemory-mcp-server. Then add the configuration to your MCP client (Claude Desktop, Cursor, etc.) and restart. The server runs locally on your machine with full offline support.
Over 113 AI clients support MCP servers, including Claude Desktop, Cursor, Windsurf, VS Code (with Cline/Continue), Zed, and many more. The Model Context Protocol is an open standard supported by the entire AI ecosystem.
Yes, the AI Memory MCP server is completely free and open-source. You can install it via pip, run it locally, and connect it to any MCP client at no cost. There are no usage limits or hidden fees.
Yes, the standalone MCP server runs entirely on your local machine. Your conversation data stays in a local SQLite database — no cloud connection required. You can also use the hosted version at aimemory.pro/api/mcp for cloud-based access.
AI Memory MCP server is a free, open-source tool focused on managing and searching existing AI conversations. Mem0 is a B2B API platform ($24M funded) for building custom memory layers in applications. AI Memory is consumer-friendly with zero setup, while Mem0 requires developer integration.
One pip install. Works with Claude, Cursor, ChatGPT, and 113+ tools.