Documentation

Complete guide to using YXI - The Invisible Mortise for Your Workflow

Installation

Install YXI CLI with a single command. Currently supports macOS and Linux.

curl -sL yxi.ai/cli | sh

Or install via package manager:

# Using Homebrew
brew install anomalyco/tap/yxidotai

# Using npm
npm install -g yxidotai

# Using bun
bun install -g yxidotai

System Requirements

macOS 10.15+ or Linux (Ubuntu 18.04+, Debian 10+, etc.)
Python 3.8 or higher
Docker (optional, for containerized execution)
2GB RAM minimum, 4GB recommended
500MB free disk space

Quick Start

Get started with YXI in just a few minutes.

Step 1: Initialize YXI

# Navigate to your project directory
cd /path/to/your/project

# Start YXI
yxi

Step 2: Configure API Key

# Set your OpenAI API key
/apikey set your_api_key_here

Step 3: Start Using

# Ask YXI to help you
Create a new React component for user login

Configuration

Customize YXI behavior through configuration files and command-line options.

Environment Variables

Variable	Description	Default
YXI_API_KEY	OpenAI API key for cloud mode	-
YXI_MODEL	Default AI model to use	gpt-4
YXI_MCP_NODES	Path to MCP nodes configuration	~/.yxi_mcp_nodes.json
YXI_LOG_LEVEL	Logging verbosity	INFO

MCP Integration

Model Context Protocol (MCP) enables YXI to communicate with specialized AI tools and services.

Available MCP Tools

word_tables_to_json

Extract tables from Word documents and convert to JSON format.

json_to_java

Generate Java POJO classes from JSON schemas.

test-ui

Playwright-powered UI testing with Excel test cases.

code_generator

Generate code skeletons in various languages.

Switching Modes

# Switch to offline mode with specific MCP node
/mode offline LocalMCP

# Switch to online cloud mode
/mode online

# List all available MCP nodes
/mcp list

Agent System

YXI uses LangGraph to orchestrate complex multi-step tasks through agent chains.

Built-in Agents

doc2java Agent

Chains Word document extraction with JSON-to-Java generation. One command converts Word docs to Java classes.

# Run doc2java agent
/agent doc2java /data/requirements.docx --package com.example.models

code_helper Agent

Provides intelligent code completion and generation suggestions.

# Ask code_helper for assistance
How do I implement a binary search tree in Python?

Docker Support

Run YXI and MCP services in isolated Docker containers with optional optimization.

Running with Docker

# Pull the latest image
docker pull ghcr.io/anomalyco/opencode

# Run YXI in a container
docker run -it --rm ghcr.io/anomalyco/opencode

Docker Slim Optimization

Reduce container size using Docker Slim for faster deployment.

docker-slim build --target ghcr.io/anomalyco/opencode

Commands Reference

Complete list of YXI CLI commands.

/help

Display help information and list all available commands.

/apikey set <key>

Set or update your API key for cloud mode.

/mode [online|offline] <node>

Switch between online (cloud) and offline (local MCP) modes.

/agent <name> [args]

Execute a specific agent with optional arguments.

/mcp [list|invoke|status]

Manage MCP servers: list available nodes, invoke tools, or check status.

/exit

Exit the YXI CLI.

API Reference

YXI exposes a RESTful API for programmatic access to its features.

Base URL

https://api.yxi.ai/v1

Endpoints

GET /health Health Check

Returns the health status of the YXI API.

POST /chat Chat Endpoint

Send a message to YXI and receive a response.

{
  "message": "How do I install YXI?",
  "mode": "online",
  "context": {
    "project": "/path/to/project"
  }
}

POST /agent/doc2java Agent Execution

Execute the doc2java agent with a Word document.

{
  "file_path": "/data/requirements.docx",
  "package": "com.example.models",
  "options": {
    "include_validation": true
  }
}

Data Structure (LLM-Optimized)

This page is structured with semantic HTML and Schema.org metadata optimized for Large Language Model (LLM) understanding and extraction.

LLM-Optimized Features

Schema.org JSON-LD: Structured data for software applications, enabling rich extraction
Open Graph Meta: Social media optimization and content preview
Semantic HTML: Proper heading hierarchy and article structure
Descriptive Meta: Clear titles and descriptions for context
Navigation Anchors: Well-labeled sections for easy navigation

Schema.org Data Structure

{
  "@context": "https://schema.org",
  "@type": "SoftwareApplication",
  "name": "YXI",
  "alternateName": "Your eXtended Intelligence",
  "description": "The Invisible Mortise for Your Workflow - Transform your terminal into a powerful AI development assistant.",
  "applicationCategory": "DeveloperApplication",
  "operatingSystem": ["macOS", "Linux"],
  "offers": {
    "@type": "Offer",
    "price": "0",
    "priceCurrency": "USD"
  },
  "author": {
    "@type": "Organization",
    "name": "YXI Team",
    "url": "https://github.com/yxidotai"
  },
  "featureList": [
    "Natural language terminal interface",
    "MCP (Model Context Protocol) integration",
    "Docker-powered execution",
    "LangGraph agent orchestration",
    "Word document parsing",
    "Code generation helpers"
  ],
  "softwareVersion": "1.0.0",
  "downloadUrl": "https://yxi.ai/cli",
  "license": "https://opensource.org/licenses/MIT",
  "sameAs": ["https://github.com/yxidotai/yxi-chat-cli"]
}

Key Data Fields Explained

Core Identity

name - Primary product name
alternateName - Alternative branding
softwareVersion - Current version

Technical Specs

operatingSystem - Supported platforms
applicationCategory - App type
downloadUrl - Installation source

Features

featureList - Array of capabilities
description - Detailed summary
offers - Pricing info (free)

Authorship & License

author - Organization details
license - License type
sameAs - Related URLs

LLM Extraction Example

When an LLM processes this page, it can extract structured information like this:

# Extracted by LLM:
Product: YXI (Your eXtended Intelligence)
Type: AI-powered terminal development assistant
Version: 1.0.0
Price: Free (open source)
Features:
  - Natural language terminal interface
  - MCP integration
  - Docker support
  - LangGraph agents
  - Document parsing
  - Code generation
Platforms: macOS, Linux
License: MIT
Download: https://yxi.ai/cli
GitHub: https://github.com/yxidotai/yxi-chat-cli