Overview
CodeCritiq is a VS Code extension that acts as your automated code review partner. Built with Groq's LLM (llama-3.3-70b-versatile) and Babel AST parsing, it:
- Analyzes your code automatically on every save
- Detects bugs, performance issues, security risks, readability problems, and architectural flaws
- Provides inline diagnostics (underlines) with severity-coded warnings
- Offers hover explanations with why-it-matters and how-to-fix instructions
- Delivers one-click Quick Fixes (Ctrl+.) with AI-generated fixed code
- Remembers past mistakes using a vector similarity store to flag repeated issues
Features
🔍 Automated Code Review on Save
Every time you save a file, the extension triggers a multi-layer review pipeline:
AST Static Analysis → Memory Recall → AI LLM Review → Merge → Display
📋 Multi-Layer Detection
| Layer |
Technology |
What It Catches |
| Static Regex |
Pattern matching |
var, console.log, TODO, empty catch, eval(), hardcoded secrets |
| AST Analysis |
Babel parser + traverse |
Long functions, empty functions, deep nesting, loose equality, any type, var |
| AI Review |
Groq llama-3.3-70b |
Bugs, logic errors, performance, security, bad practices |
| Memory |
Vector similarity (cosine) |
Repeat offenses flagged with high severity |
💡 Inline Diagnostics
Errors and warnings appear as colored underlines in your editor:
- 🔴 High severity (Error) — security risks, repeat offenses
- 🟡 Medium severity (Warning) — bugs, architecture issues
- 🔵 Low severity (Information) — style, readability tips
🖱️ Hover Explanations
Hover over any flagged line to see:
- Issue title and explanation
- Why it matters in real-world terms
- Step-by-step fix instructions
- AI-generated fixed code snippet
⚡ One-Click Quick Fixes
Hit Ctrl+. (or Cmd+. on macOS) on a flagged line to apply the AI-suggested fix automatically.
🧠 Persistent Memory
The extension stores issues in a vector database (data/memory.json) and retrieves similar past mistakes to:
- Flag repeated problems with high severity
- Inject past context into AI prompts for smarter reviews
🌐 Multi-Language Support
Active for: JavaScript, TypeScript, Python, Java, C++, C
Tech Stack
| Layer |
Technology |
| Extension Host |
VS Code API (v1.110+) |
| Language |
TypeScript (ES2022) |
| AI/LLM |
Groq SDK — llama-3.3-70b-versatile |
| AST Parsing |
Babel (@babel/parser, @babel/traverse) |
| Memory Store |
Custom vector store with cosine similarity |
| Debouncing |
Lodash |
| Environment |
dotenv |
| Build |
TypeScript Compiler (tsc) |
| Lint |
ESLint + typescript-eslint |
| Test |
Mocha (VS Code Test CLI) |
Screenshots
Code Review in Action
Hover Explanation
Quick Fix (Ctrl+.)
Diagnostics Overview
Architecture
┌─────────────────────────────────────────────────────────┐
│ VS Code Extension Host │
│ │
│ ┌──────────────────────────────────────────────────┐ │
│ │ extension.ts (Entry Point) │ │
│ │ ┌──────────────┐ ┌───────────────────────────┐ │ │
│ │ │ Save Listener │ │ Debounce Map (1.5s/file) │ │ │
│ │ └──────┬───────┘ └───────────────────────────┘ │ │
│ └─────────┼─────────────────────────────────────────┘ │
│ │ │
│ ┌─────────▼─────────────────────────────────────────┐ │
│ │ core/reviewer.ts (Pipeline) │ │
│ │ │ │
│ │ ┌──────────┐ ┌──────────┐ ┌────────────────┐ │ │
│ │ │AST │ │Memory │ │LLM Client │ │ │
│ │ │Analyzer │ │Manager │ │(Groq API) │ │ │
│ │ │(Babel) │ │(Vector │ │llama-3.3-70b │ │ │
│ │ │ │ │ Store) │ │ │ │ │
│ │ └─────┬────┘ └────┬─────┘ └───────┬────────┘ │ │
│ │ └────────────┼────────────────┘ │ │
│ │ ▼ │ │
│ │ mergeIssues() (Deduplicate) │ │
│ │ │ │ │
│ │ ▼ │ │
│ │ Store high-severity to memory │ │
│ └─────────────────────┼─────────────────────────────┘ │
│ │ │
│ ┌─────────────────────▼─────────────────────────────┐ │
│ │ Display Layer │ │
│ │ ┌────────────────┐ ┌──────────────┐ ┌───────┐ │ │
│ │ │ diagnostics.ts │ │hoverProvider │ │Code │ │ │
│ │ │ (Underlines) │ │ (Explanations)│ │Action │ │ │
│ │ └────────────────┘ └──────────────┘ │Provider│ │ │
│ │ │(Ctrl+.)│ │ │
│ │ └────────┘ │ │
│ └──────────────────────────────────────────────────┘ │
└─────────────────────────────────────────────────────────┘
Installation
Prerequisites
Steps
Clone the repository
git clone https://github.com/Krishna11-cell/ai-pair-programmer.git
cd ai-pair-programmer
Install dependencies
npm install
Configure your API key
cp .env.txt .env
Then edit .env and add your Groq API key:
GROQ_API_KEY=gsk_your_api_key_here
Compile the extension
npm run compile
Run in VS Code
- Press
F5 to open a new Extension Development Host window
- Open any supported file (
.js, .ts, .py, .java, .cpp, .c)
- Save the file to trigger a review
Usage
Triggering a Review
Reviews happen automatically on save. The extension:
- Shows
AI reviewing code... in the status bar
- After analysis, shows
AI Review complete — N issue(s) found
- Underlines issues inline with color-coded diagnostics
Viewing Issues
- Red underline = High severity (security, repeat offenses)
- Yellow underline = Medium severity (bugs, architecture)
- Blue underline = Low severity (style, readability)
Getting Fixes
- Hover over any underlined line for explanation + fix instructions
- Press
Ctrl+. (or Cmd+.) to open Quick Fix menu
- Select
Fix: ... to apply AI-generated fixed code automatically
Development
Project Structure
src/
├── extension.ts # Entry point, activation, save listener
├── diagnostics.ts # VS Code diagnostics collection manager
├── codeActionProvider.ts # Quick Fix provider (Ctrl+.)
├── analysis/
│ └── astAnalyzer.ts # Babel AST static analysis (7 rules)
├── core/
│ └── reviewer.ts # Main review pipeline orchestrator
├── hover/
│ └── hoverProvider.ts # Hover explanations from AI
├── llm/
│ └── llmClient.ts # Groq LLM API client
├── memory/
│ ├── embedding.ts # Character-code hash embeddings
│ ├── memoryManager.ts # Memory CRUD operations
│ └── vectorStore.ts # Cosine similarity vector search
└── test/
└── extension.test.ts # Mocha test suite
Scripts
| Command |
Description |
npm run compile |
Compile TypeScript to out/ |
npm run watch |
Watch mode for development |
npm run lint |
Lint all source files |
npm test |
Run VS Code extension tests |
npm run vscode:prepublish |
Build before publishing |
Roadmap
v1.0 — Current
- [x] AST static analysis (7 rules)
- [x] AI-powered code review via Groq
- [x] Inline diagnostics with severity colors
- [x] Hover explanations
- [x] One-click Quick Fixes
- [x] Persistent memory vector store
v1.1 — Upcoming
- [ ] Support for more languages (Go, Rust, Ruby)
- [ ] Custom rule configuration via VS Code settings
- [ ] Review history panel
- [ ] Performance improvements for large files
v2.0 — Future
- [ ] Code generation from natural language comments
- [ ] Team collaboration features
- [ ] CI/CD integration (GitHub Actions)
- [ ] Test generation
- [ ] Voice-controlled interactions
Contributing
Contributions are welcome! Please see CONTRIBUTING.md for guidelines.
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature)
- Commit your changes (
git commit -m 'Add amazing feature')
- Push to the branch (
git push origin feature/amazing-feature)
- Open a Pull Request
License
Distributed under the MIT License. See LICENSE for more information.
Project Link: https://github.com/Krishna11-cell/ai-pair-programmer
Built with ❤️ using TypeScript, Groq AI, and VS Code API
