mvdoc — AI-Powered Documentation Generator

Generate smart, living documentation from your Jira stories, GitHub code, and codebase — powered by Gemini AI.

Features

• Jira Integration — Pulls user stories, epics, and sprints from Jira Cloud.
• GitHub Integration — Analyzes repository structure, code, and commits.
• AI-Powered Analysis — Uses Gemini, OpenAI, Groq, or Local LLMs to generate technical specs.
• Auto Diagrams — Generates Mermaid.js flowcharts, sequence diagrams, and ER diagrams autonomously.
• Ask Your Docs (RAG) — Interact with your documentation using an AI-powered chat interface.
• VitePress Site — Produces a beautiful, searchable documentation website out of the box.

Quick Start

# Install globally
npm install -g mvdoc-cli

# Navigate to your project directory
cd my-project

# Initialize (interactive setup)
mvdoc init

# Generate documentation files
mvdoc generate

# Preview locally
mvdoc serve

# Build for production deployment
mvdoc build

Commands

mvdoc init

Interactive setup wizard that configures:

• Jira connection details (host, project key, API token)
• GitHub connection (repository, access token)
• AI Provider & Model Selection (Gemini, xAI Grok 2, Groq LLaMA 3, OpenAI GPT-4o)
• Documentation output directory preferences

This command creates .mvdocrc.json (configuration) and .env (secrets).

mvdoc generate

Runs the full documentation generation pipeline:

1. Aggregates data from Jira, GitHub, and the local codebase.
2. Processes user stories into technical specifications via AI.
3. Generates Mermaid diagrams (architecture, data flow, sequence, ER).
4. Creates Markdown pages formatted for VitePress.

# Full generation pipeline
mvdoc generate

# Skip AI processing (use raw data only)
mvdoc generate --skip-ai

# Skip diagram generation to speed up the process
mvdoc generate --skip-diagrams

# Preview what would be generated without writing to disk
mvdoc generate --dry-run

# Limit data aggregation to specific sources
mvdoc generate --source local,github

mvdoc serve

Starts a local VitePress development server to preview your generated documentation.

mvdoc serve
mvdoc serve --port 3000
mvdoc serve --no-open

mvdoc build

Compiles the VitePress site into static HTML/CSS/JS for production deployment.

mvdoc build
mvdoc build --out-dir ./public

mvdoc index

Indexes the generated documentation to enable the RAG (Retrieval-Augmented Generation) chatbot. Creates vector embeddings locally.

mvdoc index
mvdoc index --force  # Re-index all documents

mvdoc chat

Interact with your indexed documentation using AI.

# Start an interactive CLI chat session
mvdoc chat

# Start a background API server for the VitePress web widget
mvdoc chat --serve
mvdoc chat --serve --port 3456

Configuration

.mvdocrc.json

{
  "project": {
    "name": "My Project",
    "description": "A description of the project"
  },
  "sources": {
    "jira": {
      "host": "https://your-company.atlassian.net",
      "projectKey": "PROJ",
      "issueTypes": ["Story", "Epic", "Bug"]
    },
    "github": {
      "owner": "your-org",
      "repo": "your-repo",
      "branch": "main"
    },
    "local": {
      "path": ".",
      "include": ["src/**/*"],
      "exclude": ["node_modules", "dist"]
    }
  },
  "output": {
    "dir": "./docs",
    "diagrams": true,
    "modules": ["src/**/*"]
  },
  "ai": {
    "provider": "openai",
    "model": "llama3-70b-8192",
    "baseUrl": "https://api.groq.com/openai/v1"
  }
}

.env (Do NOT commit)

[email protected]
MVDOC_JIRA_TOKEN=your-jira-api-token
MVDOC_GITHUB_TOKEN=ghp_your-github-token
MVDOC_GEMINI_KEY=AIza-your-gemini-key
MVDOC_OPENAI_KEY=your-openai-or-groq-key

Generated Documentation Structure

docs/
├── index.md              # Homepage with project metrics
├── overview.md           # AI-generated project overview
├── architecture/
│   └── index.md          # Diagrams (architecture, data flow, ER)
├── stories/
│   ├── index.md          # Story index with status tracking
│   ├── proj-1.md         # Individual story with technical specifications
│   └── epic-proj-10.md   # Epic grouping view
├── modules/
│   ├── index.md          # Module index
│   └── auth.md           # AI-analyzed module documentation
└── .vitepress/
    ├── config.mts         # Auto-generated VitePress configuration
    └── theme/
        ├── index.ts       # Custom theme configuration
        ├── style.css      # Brand styling
        └── components/
            └── ChatWidget.vue  # Interactive RAG chat component

AI Model Usage

| Feature | Primary Model | Purpose | |---------|---------------|---------| | Story to Spec | Gemini/OpenAI/Groq | Translates user stories into technical specifications | | Diagram Generation | Gemini/OpenAI/Groq | Generates structural Mermaid.js syntax | | Module Analysis | Gemini/OpenAI/Groq | Reads source code and documents modules | | Project Overview | Gemini/OpenAI/Groq | Summarizes the holistic project state | | RAG Embeddings | text-embedding-004 / text-embedding-3-small | Creates vector embeddings for document search | | Ask Your Docs | Gemini/OpenAI/Groq | Processes context to answer user queries |

License

MIT © Ameen