MCP.so
Sign In

Crosmos

@crosmos-labs

Repository

About Crosmos

Persistent memory for AI agents. Give your coding assistant organizational context that compounds — search memories with hybrid retrieval, store anything with auto entity extraction, and query a living knowledge graph that gets smarter over time.

#context#ai#memory

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "crosmos-memory": {
      "command": "npx",
      "args": [
        "-y",
        "@crosmos/crosmos-mcp"
      ],
      "env": {
        "CROSMOS_API_KEY": "<YOUR_API_KEY>"
      }
    }
  }
}

Tools

4

Search memories in Crosmos Memory Engine using hybrid retrieval. Combines semantic (vector), keyword (full-text), and graph-based retrieval. Requires a space_id — call list_spaces if you don't have one yet.

Add new memories to Crosmos Memory Engine. Content is processed through an extraction pipeline that identifies entities, relationships, and creates structured knowledge graph entries. Requires a space_id — call list_spaces if you don't have one yet.

Check the health status of the Crosmos Memory Engine API

List all memory spaces owned by the authenticated user. Call this to discover available space IDs needed by search_memories and add_memory.

Overview

What is Crosmos?

Crosmos is a persistent memory layer for AI agents. It enables agents to store and retrieve organizational knowledge across sessions using a temporal knowledge graph with hybrid retrieval.

How to use Crosmos?

Install via npx @crosmos/crosmos-mcp setup which auto-detects MCP clients, or configure manually with a JSON block setting the CROSMOS_API_KEY (obtainable at console.crosmos.dev). Supported clients include Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, opencode, Cline, Roo-Cline, and Zed.

Key features of Crosmos?

  • Graph-native memory with entity and relationship linking
  • Temporal – every fact is timestamped for time‑travel queries
  • Hybrid retrieval (semantic, keyword, graph, temporal)
  • Automatic entity and relationship extraction from raw text
  • Multi‑space isolation for projects, teams, or agents
  • Four retrieval signals fused into a single ranked result

Use cases of Crosmos?

  • Store and retrieve organizational knowledge across AI agent sessions
  • Query the knowledge graph as it existed at any past point in time
  • Isolate memory by project, team, or agent using named spaces
  • Allow agents to maintain persistent context without resets
  • Enable hybrid search combining semantic meaning with graph traversal

FAQ from Crosmos

What is Crosmos and how does it differ from flat vector databases?

Crosmos uses a temporal knowledge graph that links memories as entities and relationships, not just flat vectors, enabling richer context and time

Frequently asked questions

What is Crosmos and how does it differ from flat vector databases?

Crosmos uses a temporal knowledge graph that links memories as entities and relationships, not just flat vectors, enabling richer context and time

Comments

Basic information

Category

Memory & Knowledge

Transports

stdio

Author

crosmos-labs

Submitted by

Aditya Chaudhary

More Memory & Knowledge MCP servers

MemoryMesh

CheMiguel23

A knowledge graph server that uses the Model Context Protocol (MCP) to provide structured memory persistence for AI models.

348

RAG Documentation MCP Server

hannesrudolph

An MCP server implementation that provides tools for retrieving and processing documentation through vector search, enabling AI assistants to augment their responses with relevant documentation context.

267

Memory

modelcontextprotocol

Model Context Protocol Servers

88.1K

PLUR

plur-ai

AI agents start every session with amnesia — you re-explain the project, repeat your preferences, and correct the same mistakes over and over. PLUR gives them a memory that persists. Your agent's corrections, preference

226

minutes

silverstein

Every meeting, every idea, every voice note — searchable by your AI. Open-source, privacy-first conversation memory layer.

1.3K