Memory Nexus MCP Server

How to use

To use Memory Nexus MCP, run the MCP client that talks to Memory Nexus and your agent. You initialize the MCP by starting the client with the Memory Nexus MCP package, then load the memory service into your agent’s session. You can activate Intelligence Packs to import expert knowledge on demand, and you can save, query, and relate memories to build a living knowledge graph for your agent.

How to install

Prerequisites: ensure you have Node.js installed on your system. You will use a standard Node package tool to run Memory Nexus MCP.

Install Memory Nexus MCP by running the following command in your terminal.

npx -y @memory-nexus/mcp

Configuration for MCP clients ( Claude Code / Cursor / Windsurf )

Configure your MCP client to load Memory Nexus when the agent starts. Create or edit a configuration block named mcpServers to include Memory Nexus as a server with the proper command and environment.

{
  "mcpServers": {
    "memory-nexus": {
      "command": "npx",
      "args": ["-y", "@memory-nexus/mcp"],
      "env": {
        "NEXUS_API_KEY": "mnx_your_api_key_here"
      }
    }
  }
}

Get an API key

Obtain an API key to access Memory Nexus MCP. You can start with a free trial or purchase a plan from the Nexus Marketplace.

curl -X POST https://memory-api-production-ce4d.up.railway.app/v1/trial/start \
  -H "Content-Type: application/json" \
  -d '{"email":"[email protected]"}'

Pricing or plans at a glance

Review available tiers to choose the right balance of agents, storage, and daily call limits for your use case.

| Tier | Price | Agents | Intelligence Packs | Storage | Daily Calls | |------|-------|--------|-------------------|---------|-------------| | **Trial** | Free | 1 | 1 | 10MB | 1,000 total | | Solo | $19/mo | 1 | 3 | 500MB | 10,000 | | Crew | $49/mo | 5 | 10 | 2GB | 50,000 | | Fleet | $149/mo | 25 | 50 | 10GB | 200,000 |

Tools

Memory Nexus MCP exposes a set of tools that you can call from your agent to manage memories, run semantic searches, build knowledge graphs, and control identity continuity across sessions.

How it works

Memories are stored with semantic embeddings. When you recall a memory, the system searches by meaning rather than exact keywords. Intelligence Packs provide ready-made domain expertise so your agent can answer questions with expert context. You can awaken an identity to resume a prior session and hibernate to save session state for continuity.

Example of semantic recall and packs in action: - recall("how to set up zero-downtime deploys?") -> returns expert knowledge from the installed DevOps pack.

Architecture

Memory Nexus MCP operates as a thin client. Your Agent <-> @memory-nexus/mcp <-> Memory Nexus API. All intelligence, embeddings, pattern learning, and Intelligence Packs run on Memory Nexus servers while your agent runs locally.

Environment variables

Configure the MCP client with your API key. Optional custom API URL can be provided to point to a different Memory Nexus endpoint.

NEXUS_API_KEY=mnx_your_api_key_here
NEXUS_API_URL=https://production.memory-nexus.example

What makes Memory Nexus MCP different

Semantic search, knowledge graphs, pattern learning, Intelligence Packs, awakened/hibernated identity continuity, and MCP-native integration set Memory Nexus MCP apart as a comprehensive memory and knowledge management solution for AI agents.

License

MIT

Built by

Memory Nexus - Building the intelligence marketplace for the bot economy.

remember

Store a memory with context, importance, and metadata.

recall

Semantic search across memories by meaning rather than keywords.

connect

Create relationships between memories to form knowledge graphs.

forget

Remove a memory from storage.

stats

Provide memory statistics and usage metrics.

context

Summarize session context including preferences and decisions.

awaken

Restore identity from a prior session.

hibernate

Save session state for the next awakening.

list_hats

Browse available Intelligence Packs.

wear_hat

Activate a pack to import expert knowledge into memory.