Your AI tools remember everything. Across every device.
multirys is a self-hosted MCP memory server that gives your AI tools persistent, shared context. Self-learning API integrations pull in your company data automatically—and your context follows you across devices, teams, and tools.
One MCP server, every AI assistant
Persistent, private AI memory your whole team shares—no vendor lock‑in.
Self-learning API integrations
multirys uses MCP to discover and learn your APIs automatically. Point it at a service and it figures out the schema, endpoints, and data—no manual mapping.
Context across devices & teams
Your memory follows you. Switch between your laptop, desktop, or a teammate's machine—every AI tool picks up exactly where you left off.
One-click setup
Install the server and integrate it into your AI tools in a single step. No config files, no manual wiring—it just works.
Human-in-the-loop memory
Folder‑level permissions per assistant. Every memory the AI writes waits in a review queue until a human approves it.
Self-hosted & private
Runs entirely on your infrastructure. No data leaves your network—ever.
Cut your token costs
Pre‑optimised storage and progressive disclosure mean your AI retrieves relevant snippets, not entire documents.
Works with every MCP‑compatible tool
multirys speaks the Model Context Protocol. If your AI tool supports MCP, it works with multirys.
Claude
Cursor
Copilot
ChatGPT
Hermes
OpenClaw
Manus
Windsurf
Any tool that supports the Model Context Protocol can connect to multirys—no plugins or adapters needed.
One click. Fully integrated.
Install the server and connect your AI tools in a single step.
Launch the server
One command spins up the server and auto-discovers your docs.
docker run -p 8090:8090 -v multirys-data:/data -v ./docs:/docs:ro -e MULTIRYS_SYNC_DIRS=/docs multirys/multirys
Auto-integrate with your AI tools
multirys generates the MCP config for you. Paste it into any compatible tool—or use the one-click installer to do it automatically.
Context everywhere, instantly
Every AI tool on every device shares the same persistent memory. Your context follows you across machines, teams, and tools.
Open-source core. Free for solo use.
The self-hosted MCP server is free forever under BSL 1.1. Team features start at $15/seat.
Solo
- Full source code & self‑hosting
- Local MCP server for any assistant
- Memory review & permissions
- Community support
Team
- Shared context server (Docker)
- Central permissions per assistant
- Team review queue & audit trail
- Token‑savings dashboard
- Priority support
Design Partner
- Everything in Team
- White‑glove onboarding
- Custom connectors (SharePoint, Drive, Slack)
- Direct line to the roadmap