Introduction
Memsolus is a managed memory layer for AI applications. Store, process, and retrieve context so your agents remember everything — across sessions, across agents, and at any scale.
- Persistent memory across sessions — your agents recall facts, preferences, and decisions from every past conversation
- Automatic knowledge compilation — related memories are grouped and compiled into structured, searchable knowledge
- Hybrid vector search — semantic and keyword retrieval combined for high-precision results
- Multi-agent support — multiple agents can share the same workspace to collaborate using a common memory space
How It Works
Store memories via API, SDK, or MCP. Memsolus automatically processes them — extracting facts, removing duplicates, building a knowledge graph, and compiling related memories into structured knowledge. Search across everything with hybrid vector search.
Explore the Docs
Quick Start
Store and retrieve your first memory in under 5 minutes.
→lightbulbConcepts
Understand the memory lifecycle, knowledge base, and search.
→apiAPI Reference
Full REST API documentation with request and response examples.
→codeSDK
TypeScript SDK for Node.js and Bun environments.
→smart_toyMCP Server
Use Memsolus directly from Claude Desktop and other MCP agents.
→auto_storiesUse Cases
Real-world examples with complete working code.
→