Nexus AI Code Assist Vscode

Nexus AI Code Assist Vscode — bringing AI capabilities directly into the VS Code development environment.

---

Topics: vscode-extension · developer-tools · ai-code-completion · coding-assistant · deep-learning · generative-ai · ide-integration · large-language-models · llm · typescript

Overview

Nexus AI Code Assist Vscode is a VS Code extension that integrates AI-powered tooling into the editor workflow, eliminating context switching between the editor and external AI tools. It provides completions, explanations, generation commands, or chat capabilities directly accessible from the Command Palette and keyboard shortcuts.

The extension is built with TypeScript using the VS Code Extension API and communicates with an AI backend (Google Gemini, OpenAI, or a local model via Ollama) to provide intelligent, context-aware assistance. All sensitive credentials are stored in VS Code's SecretStorage to prevent accidental exposure.

The architecture follows the VS Code provider model: language-specific providers handle completions and hover actions, while WebView panels provide richer chat and generation interfaces. Configuration is fully exposed through VS Code's settings system for a native user experience.

---

Motivation

The best developer tools are the ones that stay out of the way while making hard things easier. This extension was built to make AI assistance feel native to the editor rather than bolted on — accessible with a keystroke, configured through familiar settings, and respectful of the developer's focus.

---

Architecture

User Action (keystroke / command / hover)
        │
  VS Code Provider (InlineCompletion / Hover / Command)
        │
  Context extraction from active editor
        │
  AI Backend API call
        │
  Editor injection / WebView display

---

Features

Command Palette Integration

All extension actions are registered as named VS Code commands, accessible from the Command Palette (Ctrl+Shift+P) without memorising keyboard shortcuts.

Inline Code Actions

Right-click context menu items for common AI actions: Explain Selection, Refactor Selection, Add Docstring, Generate Tests.

Configurable AI Backend

Switch between Google Gemini, OpenAI, or local Ollama endpoints via VS Code Settings without reinstalling the extension.

SecretStorage for API Keys

API keys are stored in VS Code's encrypted SecretStorage rather than plaintext settings, preventing accidental exposure in dotfiles or version control.

WebView Panel Interface

Rich HTML/CSS/JS panel for chat, generation history, and multi-turn interaction, rendered in VS Code's sidebar or editor area.

Status Bar Integration

Real-time status bar item shows the current AI model name and a loading indicator during active API calls.

Keyboard Shortcut Bindings

All primary actions have configurable keyboard shortcuts defined in package.json contributes.keybindings.

Cross-Platform Support

Works on VS Code for Windows, macOS, and Linux; tested on VS Code 1.85+ and compatible with VS Code Insiders.

---

Tech Stack

Library / Tool	Role	Why This Choice
TypeScript	Extension language	Type-safe VS Code extension development
VS Code Extension API	Editor integration	Providers, commands, WebView, SecretStorage
esbuild	Bundler	Fast TypeScript → JavaScript bundling
AI SDK (Gemini / OpenAI)	AI backend	API client for LLM inference
vscode-test	Testing	Extension integration tests in an Extension Host

Key packages detected in this repo: node-fetch

---

Getting Started

Prerequisites

• Python 3.9+ (or Node.js 18+ for TypeScript/JS projects)
• pip or npm package manager
• Relevant API keys (see Configuration section)

Installation

git clone https://github.com/Devanik21/nexus-ai-code-assist-vscode.git
cd nexus-ai-code-assist-vscode
npm install
npm run watch   # watch mode for development
# Press F5 in VS Code to open Extension Development Host

# Build distributable VSIX
npm run package
code --install-extension *.vsix

---

Usage

// settings.json
{
  "extension.apiKey": "",
  "extension.model": "gemini-2.0-flash",
  "extension.triggerMode": "onDemand"
}

---

Configuration

Variable	Default	Description
extension.apiKey	(empty)	AI backend API key (use SecretStorage command to set)
extension.model	gemini-2.0-flash	AI model identifier
extension.triggerMode	onDemand	When to trigger AI: onType, onDemand, onSave

Copy .env.example to .env and populate all required values before running.

---

Project Structure

nexus-ai-code-assist-vscode/
├── README.md
├── package-lock.json
├── package.json
├── tsconfig.json
└── ...

---

Roadmap

• Repository-wide context indexing for project-aware completions
• Offline local model support via Ollama REST API
• Test generation from function signatures
• Multi-file refactoring with diff preview
• Marketplace publication with automated release workflow

---

Contributing

Contributions, issues, and feature requests are welcome. Please:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/your-feature)
3. Commit your changes (git commit -m 'feat: add your feature')
4. Push to your branch (git push origin feature/your-feature)
5. Open a Pull Request

Please follow conventional commit messages and ensure any new code is documented.

---

Notes

A valid API key for the configured AI backend is required. The extension only sends code to the AI backend when a user action explicitly triggers it.

---

Author

Devanik Debnath
B.Tech, Electronics & Communication Engineering
National Institute of Technology Agartala

---

License

This project is open source and available under the MIT License.

---

Crafted with curiosity, precision, and a belief that good software is worth building well.