Cortex AI — MCP Server

MCP (Model Context Protocol) server for Cortex AI, th state-of-the-art agentic memory. Provides tools for storing, recalling, and managing memories with knowledge-graph enriched context.

Available Tools

cortex_search

Search through Cortex AI memories. Returns relevant chunks with graph-enriched context including entity paths and knowledge graph relations.

cortex_store

Save important information to Cortex AI memory. Cortex automatically extracts insights, preferences, and builds a knowledge graph from the stored content.

cortex_ingest_conversation

Ingest user-assistant conversation turns into Cortex AI memory. Cortex extracts insights, preferences, and knowledge graph entities from the conversation.

cortex_list_memories

List all stored user memories in Cortex AI. Returns memory IDs and their content.

cortex_delete_memory

Delete a specific user memory from Cortex AI by its memory ID.

cortex_fetch_content

Fetch the full content of a specific source by its source ID.

cortex_list_sources

List all ingested sources in Cortex AI memory.

Configuration

Get Your Credentials

1. Get your Cortex API Key from Cortex AI
2. Get your Tenant ID from the Cortex dashboard

Environment Variables

| Variable | Description | Default | | ------------------------ | ----------------------------------- | -------------- | | CORTEX_API_KEY | Your Cortex API key | Required | | CORTEX_TENANT_ID | Your Cortex tenant identifier | Required | | CORTEX_SUB_TENANT_ID | Sub-tenant for data partitioning | cortex-mcp | | CORTEX_LOG_LEVEL | Log level: DEBUG, INFO, WARN, ERROR | ERROR |

Claude Desktop

{
  "mcpServers": {
    "cortex": {
      "command": "npx",
      "args": ["-y", "@usecortex_ai/[email protected]"],
      "env": {
        "CORTEX_API_KEY": "your-api-key",
        "CORTEX_TENANT_ID": "your-tenant-id"
      }
    }
  }
}

Cursor & Windsurf

| Client | Config File | | -------- | --------------------------------------- | | Cursor | ~/.cursor/mcp.json | | Windsurf | ~/.codeium/windsurf/mcp_config.json |

{
  "mcpServers": {
    "cortex": {
      "command": "npx",
      "args": ["-y", "@usecortex_ai/[email protected]"],
      "env": {
        "CORTEX_API_KEY": "your-api-key",
        "CORTEX_TENANT_ID": "your-tenant-id"
      }
    }
  }
}

VS Code

Add to .vscode/mcp.json:

{
  "servers": {
    "cortex": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@usecortex_ai/[email protected]"],
      "env": {
        "CORTEX_API_KEY": "your-api-key",
        "CORTEX_TENANT_ID": "your-tenant-id"
      }
    }
  }
}

Custom Sub-Tenant

To partition data, set the CORTEX_SUB_TENANT_ID environment variable:

{
  "mcpServers": {
    "cortex": {
      "command": "npx",
      "args": ["-y", "@usecortex_ai/[email protected]"],
      "env": {
        "CORTEX_API_KEY": "your-api-key",
        "CORTEX_TENANT_ID": "your-tenant-id",
        "CORTEX_SUB_TENANT_ID": "my-project"
      }
    }
  }
}

How It Works

• cortex_search queries /recall/recall_preferences for relevant memories and returns graph-enriched context (entity paths, chunk relations, extra context).
• cortex_store sends text to /memories/add_memory with infer: true and upsert: true. Cortex extracts insights and builds a knowledge graph automatically.
• cortex_ingest_conversation sends user-assistant pairs to /memories/add_memory as conversation turns, grouped by source_id.
• cortex_list_memories and cortex_list_sources query /list/data to browse stored data.
• cortex_delete_memory calls DELETE /memories/delete_memory to remove a specific memory.
• cortex_fetch_content calls /fetch/content to retrieve the original ingested content.

Development

npm install
npm run build
CORTEX_API_KEY=your-key CORTEX_TENANT_ID=your-tenant npm start

For development with auto-reload:

CORTEX_API_KEY=your-key CORTEX_TENANT_ID=your-tenant npm run dev

Troubleshooting

• API Key Issues: Ensure CORTEX_API_KEY is set correctly
• Connection Errors: Check your internet connection and API key validity
• Tool Not Found: Make sure the package is installed and the command path is correct
• Debug Logging: Set CORTEX_LOG_LEVEL=DEBUG for verbose output