CyberMem
Your AI Memory — Install Everywhere
60-Second Production Deploy
Install CyberMem
In your terminal, run:
npx @cybermem/cli initCyberMem will auto generate Docker Compose for local usage
Open Dashboard
The dashboard URL will be shown after successful deployment. Open it to setup your clients.
A Security Token is generated and shown in the dashboard. Keep it secret.
See setup guideConnect Your Clients
Use the MCP configuration provided in the dashboard to connect Claude, Antigravity, Gemini, and other AI assistants.
Enjoy! 🎉
Your AI assistants now share a unified memory. They remember context across conversations and devices.
Powerful Capabilities
Everything you need to build intelligent, memory-aware AI agents.
100% Private. Always.
CyberMem never uploads your data. All memory is stored locally in SQLite. Metrics are handled via an embedded, low-footprint exporter. Your secrets stay on your hardware.
Security by Environment
🏠 Local
No HTTPS needed. Keyless access for localhost.
🍓 Raspberry Pi
Use --remote-access for Tailscale HTTPS.
☁️ Cloud/VPS
Auto-SSL via Caddy / Traefik. API key required.
Full Data Control
Your box, your data. No remote telemetry. You own the database file.
No Monthly Fees
Self-hosted = Free forever. No SaaS subscription needed. Just bring your own API key for LLMs.
Production-Grade Architecture
Not just a script. A complete observability and security pipeline.
Built with Platform Engineering best practices: Infrastructure as Code, Zero-Trust Security, Observability-First Design, and Production-Ready Automation.
AI Clients
Dashboard
Traefik Gateway
Auth Sidecar
- Auto-HTTPS
- API Key Auth
- Client ID Extraction
Vector Pipeline
Log Processor
- Zero-overhead
- Logs to Metrics
- Prometheus Exporter
OpenMemory Core
AI Memory Engine
Private Vault
Your Hardware
Comparison / Alternatives
CyberMem combines the privacy of local hosting with the features of cloud platforms.
| Feature | CyberMem | Mem0 | OpenMemory | Supermemory | Official MCP |
|---|---|---|---|---|---|
| SaaS / Managed Cloud | Private Only | Available | - | Core Product | - |
| Self-Hosted (RPi / Local) | Native Support | No Docs | Local-first | SaaS Only | Reference |
| Self-Hosted (VPS) | Docker Optimized | No Docs | Manual | SaaS Only | Local only |
| Easy Setup & Install | 1-Command | Complex | Manual | Account | Config Heavy |
| Infrastructure Generation | Automated (Docker/K8s) | - | - | - | - |
| Knowledge Graph Engine | Temporal | Mem0g | Temporal | Graph | Basic |
| Per-Client Monitoring | Built-in | - | - | - | - |
| Full Request Audit | Per-Client | - | - | - | - |
Deploy Everywhere. Equally Simple.
3 Platforms • 1 Interface • 100% Data Portability
Local
Simple as Homebrew
Raspberry Pi
Like Homebridge / Outline
AWS/Cloud
K8s as a 1-Click App
Visual Dashboard
Inspect, debug, and operate AI memory state.
Real-time Metrics
Track latency, operations per second, and memory usage
Per-client Monitoring
Inspect memory state and performance for each agent separately
Easy Configuration
Manage API keys, retention policies, and integrations
Common Questions
What is MCP?
The Model Context Protocol (MCP) is an open standard that enables AI models to securely access data and tools.
CyberMem is your shared AI memory server.
It allows different AI tools (Claude, Cursor, etc.) to share a single, persistent long-term memory.
How can I transfer my data?
Your memories are stored in a portable SQLite file that you fully own and control.
Create a backup:
npx @cybermem/cli backupRestore from backup:
npx @cybermem/cli restore ./backup.db