rtlviz 0.9.0

MCP Server for AI-powered RTL diagram image generation via Gemini

RTLViz

Generate interactive block diagrams from Verilog source files.

Installation

Step 1: Install Package

pip install rtlviz

Step 2: Setup IDE

Spins up the MCP server config for all popular IDEs.

rtlviz setup

Step 3: Manual IDE Setup (if Step 2 fails)

If the automatic setup doesn't work, you can manually configure your IDE's MCP settings (e.g., claude_desktop_config.json).

Command: rtlviz-server works if installed in PATH. Otherwise use python -m rtlviz.

JSON Config:

{
  "mcpServers": {
    "rtlviz": {
      "command": "python",
      "args": ["-m", "rtlviz"]
    }
  }
}

Usage

AI Agent

After setup, ask your AI assistant (Role: Software Engineer):

"Use rtlviz MCP to generate a diagram for the Verilog files in <path to code>"

The diagram will automatically open in your browser.

Recommended Models: For best results, use state-of-the-art reasoning models:

• Claude: Opus 4.5 , Sonnet 4.5
• Google: Gemini 3
• OpenAI: GPT-5.2

Example Output

Features

• Automatic Parsing - Extractions modules and connections from Verilog
• Pipeline Detection - Identifies CPU stages (IF, ID, EX, MEM, WB)
• Beautiful Diagrams - Orthogonal routing with semantic clustering
• Interactive - Pan, zoom, and export to SVG
• Auto-Open - Instantly opens generated diagrams in browser

Requirements

• Python 3.10+
• Internet connection (for Viz.js rendering)

Authors

• Naveen Venkat ([e-mail])
• Hari Ayyappane ([e-mail])

License

MIT License