# CodeCortex Context Engine MCP server

> Provides context intelligence for AI coding agents by mapping repositories, resolving symbols, and estimating the impact of code changes.

- Repository: https://github.com/BehnamJalaliCo/CodeCortex
- Page: https://hostedmcp.org/servers/behnamjalalico/codecortex
- Transport: stdio
- Stars: 8
- License: Apache-2.0
- Latest version: 0.1.0a10

## Install

### Claude Code

```sh
claude mcp add codecortex -- uvx codecortex-context-engine
```

### Cursor (`~/.cursor/mcp.json`)

```json
{
  "mcpServers": {
    "codecortex": {
      "command": "uvx",
      "args": [
        "codecortex-context-engine"
      ]
    }
  }
}
```

### VS Code (`.vscode/mcp.json`)

```json
{
  "servers": {
    "codecortex": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "codecortex-context-engine"
      ]
    }
  }
}
```

### VS Code CLI

```sh
code --add-mcp '{"name":"codecortex","type":"stdio","command":"uvx","args":["codecortex-context-engine"]}'
```

### Claude Desktop (`claude_desktop_config.json`)

```json
{
  "mcpServers": {
    "codecortex": {
      "command": "uvx",
      "args": [
        "codecortex-context-engine"
      ]
    }
  }
}
```

### Codex

```sh
codex mcp add codecortex -- uvx codecortex-context-engine
```

### Codex config.toml (`~/.codex/config.toml`)

```toml
[mcp_servers.codecortex]
command = "uvx"
args = ["codecortex-context-engine"]
```

### Gemini CLI

```sh
gemini mcp add -s user codecortex uvx -- codecortex-context-engine
```

### Gemini settings.json (`~/.gemini/settings.json`)

```json
{
  "mcpServers": {
    "codecortex": {
      "command": "uvx",
      "args": [
        "codecortex-context-engine"
      ]
    }
  }
}
```

## FAQ

### What does CodeCortex do for AI coding agents?

It turns a software repository into a query-specific evidence system, providing intelligence about structure, symbols, relationships, and Git history.

### Which coding agents are supported?

It includes merge-safe project configuration for Claude Code, Codex, Cursor, Gemini CLI, and OpenCode.

### How can I expose my repository to an MCP-capable agent?

You can use the command `cortex mcp --path .` to expose the repository.

### What are the system requirements?

The engine requires Python versions 3.11 through 3.13.