AI README Generator
An enterprise-grade VS Code extension that automatically scans your codebase, analyzes its structure, and uses state-of-the-art NVIDIA-hosted open-source LLMs (like Llama 3.3 and Nemotron) to generate beautiful, comprehensive, and professional README.md files.
Featuring an interactive, VS Code-themed sidebar Webview UI, the extension provides workspace diagnostics, section-by-section customization, and a targeted Section Editor that lets you regenerate single parts of your documentation on-the-fly.
Key Features
- 🔍 Smart Workspace Analysis: Parses languages, package managers, dependency trees, environment configuration files (
.env.example), Docker configs, CI/CD scripts, and git remotes.
- ⚡ NVIDIA NIM Integration: Leverages NVIDIA's high-speed API platform for low-latency Markdown generation using advanced open-source models like
meta/llama-3.3-70b-instruct.
- 🎨 Native VS Code Theme Integration: Styled using CSS variables (
--vscode-*), adapting perfectly to dark, light, or high-contrast themes.
- 🛠️ Section Checklist & Custom Prompts: Toggle specific sections (Installation, Usage, Folder Structure, API, FAQ, etc.) and supply custom markdown rules or styling guidelines.
- 📝 Live Markdown Preview: Render and read your generated README live using
markdown-it, or switch to Raw text edit mode.
- 🔄 Targeted Section Regeneration: Highlight or select individual sections, supply direct prompts (e.g. "Add database migration steps"), and let the LLM rebuild just that section without touching the rest of your README.
- 📁 Project Explorer Dashboard: A split-pane dashboard that dynamically renders collapsible recursive Tree Views (left) and Data Tables (right) from JSON or XML files.
- 🔐 Secure Key Storage: Uses VS Code's native
SecretStorage to save your NVIDIA API key securely.
Tech Stack
- Extension Framework: VS Code Extension API
- Language: TypeScript
- Frontend UI: React
- Build Tools: Vite (for React Webview) & Esbuild (for Extension Host)
- AI Requests: Axios (pointing to NVIDIA's completions endpoint)
- Markdown Rendering:
markdown-it
- File System Parsing: Node
fs/promises & path APIs
- XML Parsing:
xml2js (translates XML tree nodes and rows to JSON)
- Icons:
lucide-react
Project Structure
Ai_README_Generator/
├── .vscode/
│ ├── launch.json # Run Extension configuration profile
│ └── tasks.json # Compiles extension & webview in sync
├── data/
│ ├── sample-data.json # Baseline JSON explorer data
│ └── sample-data.xml # Baseline XML explorer data
├── media/
│ └── icon.svg # Extension sidebar logo
├── out/ # Compiled assets (bundled via esbuild and Vite)
│ ├── extension.js # Extension host bundle
│ └── webview/ # Compiled React app
│ ├── index.html
│ └── assets/
│ ├── index.js
│ └── index.css
├── src/ # Extension backend
│ ├── analyzer.ts # Workspace scanning and metadata extraction
│ ├── extension.ts # Activates commands and registers Webview
│ ├── generator.ts # Queries NVIDIA NIM completions API
│ └── webviewPanel.ts # Orchestrates IPC bridge, JSON/XML parses & SecretStorage
├── webview-ui/ # React Webview client code
│ ├── src/
│ │ ├── App.tsx # Main UI state, panels, and tabs
│ │ ├── index.css # Glassmorphic, responsive stylesheets
│ │ └── main.tsx # Mounting script
│ ├── index.html
│ ├── tsconfig.json
│ └── vite.config.ts # Maps outDir to parent output folder
├── package.json # Extension manifest
├── tsconfig.json # Extension backend compiler options
└── esbuild.js # Bundler for extension backend
Configuration Settings
You can customize the extension via VS Code settings (ctrl+, or cmd+,):
| Setting |
Type |
Default Value |
Description |
aiReadmeGenerator.defaultModel |
string |
meta/llama-3.3-70b-instruct |
The default NVIDIA NIM model identifier. |
aiReadmeGenerator.temperature |
number |
0.5 |
Sampling temperature. Higher values mean more creative outputs. |
aiReadmeGenerator.maxTokens |
integer |
4096 |
Maximum token length for generated outputs. |
Installation & Setup
VsCode MarketPlace link
Prerequisites
- VS Code: Version
1.74.0 or higher.
- NVIDIA API Key: Register and obtain a free API key at the NVIDIA API Catalog.
Installation
Method 1: From Visual Studio Marketplace
- Open VS Code.
- Go to the Extensions View (
Ctrl+Shift+X or Cmd+Shift+X).
- Search for AI README Markdown Generator.
- Click Install.
Method 2: Manual VSIX Installation
If you downloaded the .vsix bundle directly (e.g., ai-readme-markdown-generator-1.0.5.vsix):
- Open VS Code.
- Open the Extensions View, click the
... (More Actions) button in the top-right corner of the Extensions panel.
- Select Install from VSIX....
- Choose the downloaded
.vsix file and click Install.
Contributing & Development
If you wish to build, run, or contribute to the extension source code:
- Clone the repository:
git clone https://github.com/NEO-D-2004/AI-README-Generator.git
cd AI-README-Generator
- Install Dependencies:
Install dependencies for the extension and front-end:
npm install
- Compile and Build:
Build the backend files and React webview assets simultaneously:
npm run build-all
- Run and Debug:
- Open the folder in VS Code.
- Press
F5 (or go to Run & Debug and click Run Extension).
- In the new Extension Development Host window, click the AI README icon in the sidebar Activity Bar.
- Input your NVIDIA API key and start developing!
Usage Guide
- Scan Workspace: When you open the extension panel in a project, it automatically performs a framework, language, and folder tree scan. Click Rescan Workspace at any time to refresh the metrics.
- Configure Settings:
- Choose a model (e.g., Llama 3.3 70B, Llama 3.1 8B, or Nemotron).
- Adjust generation temperature and token constraints.
- Toggle options for badges and diagrams.
- Select Sections: Check or uncheck sections in the README Sections checklist to include only the items you want.
- Generate README: Click Generate README.md. The live progress loader will show generation status.
- Review & Edit:
- Preview: View the parsed markdown rendered in real-time.
- Raw: Edit the raw markdown text block directly.
- Section Editor: Select an isolated section (e.g., Installation), write a instruction (e.g. "add yarn global install instruction"), and click Regenerate Section. The editor will query the AI and replace only that section in your draft.
- Save: Click the Save button in the tab bar. This writes the completed
README.md to your workspace root and automatically opens it in your editor.
- Explorer Dashboard: Click the Explorer Dashboard tab in the webview. Toggle between the dropdown formats (
sample-data.json / sample-data.xml) to see the recursive Tree View (collapsible directories, folder/file icons) and the corresponding file statistics Data Table. Click Refresh to reload the files if they are modified.
License
This project is licensed under the MIT License.
| |
