# Rowih MCP server

> Enables AI agents to communicate using a structured language that explicitly states the force, basis, and authority of every message move.

- Repository: https://github.com/rowih/rowih
- Page: https://hostedmcp.org/servers/rowih/rowih
- Transport: stdio
- Stars: 1
- License: Apache-2.0
- Latest version: 1.1.8

## Install

### Claude Code

```sh
claude mcp add rowih -- uvx rowih
```

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

```json
{
  "mcpServers": {
    "rowih": {
      "command": "uvx",
      "args": [
        "rowih"
      ]
    }
  }
}
```

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

```json
{
  "servers": {
    "rowih": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "rowih"
      ]
    }
  }
}
```

### VS Code CLI

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

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

```json
{
  "mcpServers": {
    "rowih": {
      "command": "uvx",
      "args": [
        "rowih"
      ]
    }
  }
}
```

### Codex

```sh
codex mcp add rowih -- uvx rowih
```

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

```toml
[mcp_servers.rowih]
command = "uvx"
args = ["rowih"]
```

### Gemini CLI

```sh
gemini mcp add -s user rowih uvx -- rowih
```

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

```json
{
  "mcpServers": {
    "rowih": {
      "command": "uvx",
      "args": [
        "rowih"
      ]
    }
  }
}
```

## FAQ

### What is Rowih?

Rowih is a language designed specifically for AI agents to communicate with one another, detailing the intent, foundation, and authority of each message.

### How does the language ensure message integrity?

The toolchain uses check words for word counts and moves, and an integrity check value to detect any edits to symbols or words.

### Can agents audit their interactions?

Yes, requests, commitments, and their withdrawals are recorded as entries that each participating party can keep and audit.
