Cortex MCP

Multi-level reasoning MCP server with configurable depth levels, session-based state management, structured thought input, and real-time trace resources.

Overview

Cortex MCP is a stdio-only MCP server for stateful, depth-controlled reasoning. The runtime entrypoint in src/index.ts connects createServer() to StdioServerTransport, and the server surface in src/server.ts enables tools, prompts, completions, logging, and subscribable resources around a single session-based reasoning engine.

The live MCP surface confirmed by Inspector is 1 tool, 6 concrete resources, 4 resource templates, and 7 prompts. Sessions are stored in memory, exposed as MCP resources, and cleared on process restart.

Key Features

• reasoning_think supports step-by-step sessions, run_to_completion batches, rollback, early conclusion, and structured observation / hypothesis / evaluation input.
• Four depth levels are built into the engine: basic, normal, high, and expert, each with bounded thought ranges and token budgets.
• Prompt helpers expose reasoning.basic, reasoning.normal, reasoning.high, reasoning.expert, reasoning.continue, reasoning.retry, and get-help.
• Resource endpoints expose internal docs plus live session lists, per-session JSON views, full markdown traces, and individual thought documents.
• Completions are wired for levels, session IDs, and thought names through completable() and resource-template completion hooks.

Requirements

• Node.js >=24 for local npx or npm usage.
• An MCP client that supports stdio transport.
• Optional: Docker if you want to build or run the container image defined by Dockerfile.

Quick Start

Use this standard MCP client configuration:

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

Client Configuration

{
  "servers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

code --add-mcp '{"name":"cortex-mcp","command":"npx","args":["-y","@j0hanz/cortex-mcp@latest"]}'

{
  "servers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

code-insiders --add-mcp '{"name":"cortex-mcp","command":"npx","args":["-y","@j0hanz/cortex-mcp@latest"]}'

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

claude mcp add cortex-mcp -- npx -y @j0hanz/cortex-mcp@latest

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

amp mcp add cortex-mcp -- npx -y @j0hanz/cortex-mcp@latest

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "servers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "context_servers": {
    "cortex-mcp": {
      "settings": {
        "command": "npx",
        "args": ["-y", "@j0hanz/cortex-mcp@latest"]
      }
    }
  }
}

{
  "augment.advanced": {
    "mcpServers": [
      {
        "id": "cortex-mcp",
        "command": "npx",
        "args": ["-y", "@j0hanz/cortex-mcp@latest"]
      }
    ]
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

{
  "mcpServers": {
    "cortex-mcp": {
      "command": "npx",
      "args": ["-y", "@j0hanz/cortex-mcp@latest"]
    }
  }
}

Use Cases

Start bounded reasoning at the right depth

Use reasoning.basic, reasoning.normal, reasoning.high, or reasoning.expert when the client wants a prompt-first entrypoint, or call reasoning_think directly with query, level, and the first thought. Each response returns the current session state plus a summary string that tells the client how to continue.

Relevant tool: reasoning_think
Related prompts: reasoning.basic, reasoning.normal, reasoning.high, reasoning.expert

Continue, retry, or batch an active session

Reuse sessionId to continue a prior trace, switch to runMode="run_to_completion" when you already have the remaining thought inputs, or use the continuation and retry prompts to generate the next call payload. The handler also supports rollbackToStep and isConclusion for revising or ending a trace early.

Relevant tool: reasoning_think
Related prompts: reasoning.continue, reasoning.retry

Inspect live traces without re-running the tool

Read reasoning://sessions for the active session list, reasoning://sessions/{sessionId} for the JSON detail view, reasoning://sessions/{sessionId}/trace for the markdown transcript, or reasoning://sessions/{sessionId}/thoughts/{thoughtName} for a single thought. This lets a client present progress or audit a session independently from the next tool call.

Relevant resources: reasoning://sessions, reasoning://sessions/{sessionId}, reasoning://sessions/{sessionId}/trace, reasoning://sessions/{sessionId}/thoughts/{thoughtName}

Architecture

[MCP Client]
    |
    | stdio
    v
[src/index.ts]
    createServer()
    -> new StdioServerTransport()
    -> server.connect(transport)
    |
    v
[src/server.ts]
    McpServer("cortex-mcp")
    capabilities:
      - tools
      - prompts
      - completions
      - logging
      - resources { subscribe: true, listChanged: true }
    |
    +--> tools/call
    |     -> reasoning_think
    |     -> src/tools/reasoning-think.ts
    |     -> ReasoningThinkInputSchema / ReasoningThinkToolOutputSchema
    |     -> src/engine/reasoner.ts
    |     -> SessionStore
    |
    +--> prompts/get
    |     -> src/prompts/index.ts
    |
    +--> resources/read
    |     -> src/resources/index.ts
    |     -> internal://* and reasoning://sessions/*
    |
    +--> notifications
          -> logging messages
          -> resources/list_changed
          -> resources/updated
          -> notifications/progress

Request Lifecycle

[Client] -- initialize --> [Server]
[Server] -- serverInfo + capabilities --> [Client]
[Client] -- notifications/initialized --> [Server]
[Client] -- tools/call {name: "reasoning_think", arguments} --> [Handler]
[Handler] -- validate args --> [Reasoner + SessionStore]
[Reasoner] -- progress/resource events --> [Server notifications]
[Handler] -- structuredContent + optional trace resource --> [Client]

MCP Surface

Tools

reasoning_think

Stateful reasoning tool for creating and continuing multi-step sessions. It supports one-step interactive calls, run_to_completion batches, structured observation/hypothesis/evaluation input, rollback, and early conclusion while returning structured session state.

1. [Client] -- tools/call {name: "reasoning_think", arguments} --> [Server]
   Transport: stdio
2. [Server] -- dispatch("reasoning_think") --> [Handler: src/tools/reasoning-think.ts]
3. [Handler] -- validate(ReasoningThinkInputSchema) --> [src/engine/reasoner.ts]
4. [Reasoner] -- create/update session --> [src/engine/session-store.ts]
5. [Handler] -- structuredContent + optional embedded trace resource --> [Client]

Resources

Prompts

MCP Capabilities

Tool Annotations

Structured Output

• reasoning_think declares outputSchema and returns structuredContent, with an embedded trace resource when the trace is small enough. Evidence: src/tools/reasoning-think.ts:498, src/lib/mcp.ts:97-114.

Configuration

[!NOTE] The source does not define any HTTP host/port configuration. The only other environment-related signal is NODE_ENV=production in the Docker image and --env-file=.env in the local dev:run script.

Security

[!NOTE] No auth, OAuth, HTTP origin checks, or rate-limiting controls are implemented in the current source because the server only exposes stdio transport.

Development

Additional helper scripts for diagnostics, coverage, asset copying, and knip are defined in package.json.

Build and Release

• .github/workflows/release.yml bumps package.json and server.json, then runs npm run lint, npm run type-check, npm run test, and npm run build before tagging and creating a GitHub release.
• The same workflow publishes the package to npm with Trusted Publishing, publishes to the MCP Registry with mcp-publisher, and pushes a multi-arch Docker image to ghcr.io.
• Dockerfile uses a multi-stage Node 24 Alpine build, prunes dev dependencies, and runs the released container as the mcp user.

Troubleshooting

• Sessions are in memory and expire after 30 minutes by default. If you receive E_SESSION_NOT_FOUND, start a new session or increase CORTEX_SESSION_TTL_MS.
• runMode="run_to_completion" requires enough thought entries to cover the remaining steps. If you want the server to return after each step, keep the default step mode.
• For stdio transport, do not add custom stdout logging around the server process. This server routes logs through MCP logging and falls back to stderr on failures.

Credits

Contributing and License

• License: MIT
• Contributions are welcome via pull requests.