SieveAi

Advanced AI-Powered Code Review Tool with hybrid static + Active AI analysis

📦 Available on npm: npm install -g sieveai | View on NPM

SieveAi is an intelligent code review tool that combines static analysis with AI-powered insights to help maintain code quality, security, and performance. It supports 15+ programming languages and provides deep analysis including const correctness, security vulnerabilities, logic errors, and architectural improvements.

Features

AI-Powered Analysis

• Logic & Bug Detection: Identifies complex logic errors, race conditions, memory leaks
• Security Analysis: Detects injection vulnerabilities, authentication issues, unsafe patterns
• Performance Optimization: Suggests const correctness, inline functions, algorithmic improvements
• Architecture Review: SOLID principles, encapsulation, design patterns

Multi-Language Support

• Systems Languages: C++, C, Rust, Go
• JVM Languages: Java, Kotlin, Scala
• Web Technologies: JavaScript, TypeScript, HTML, CSS
• Scripting: Python, Ruby, Shell, PHP
• Data Formats: JSON, YAML, Markdown
• And more: Automatic language detection

Comprehensive Analysis

• Static Analysis: Syntax checking, secret scanning, dependency analysis
• Whitespace Analysis: Formatting consistency, unnecessary diff noise detection
• Security Scanning: Built-in secret detection, vulnerability patterns
• Performance Insights: Optimization opportunities, algorithmic improvements

Professional Output

• Clean TUI Interface: Professional bordered output with severity indicators
• Multiple Formats: TUI, JSON, plain text
• Detailed Reports: Confidence scores, rationale, fix suggestions
• Git Integration: Works with diffs, staged changes, commits

Quick Start

Installation

Option 1: NPM Package (Recommended)

# Install globally for CLI usage
npm install -g sieveai

# Or install locally in your project
npm install --save-dev sieveai

Option 2: From Source

# Clone the repository
git clone https://github.com/Jayesh-Dev21/SieveAi.git
cd SieveAi

# Install dependencies
npm install

# Build the project
npm run build

# Make globally available (optional)
npm link

Setup LLM Backend

SieveAi requires an LLM for AI-powered analysis. Currently supports Ollama:

# Install Ollama (https://ollama.ai)
curl -fsSL https://ollama.ai/install.sh | sh

# Pull a supported model
ollama pull gemma3:latest
# or
ollama pull llama3.2:latest
# or
ollama pull deepseek-coder:latest

Basic Usage

# Run code review on current git changes
sieveai check

# Run with verbose output
sieveai check --verbose

# Run with specific format
sieveai check --format json

# Run with custom model
sieveai check --model ollama:deepseek-coder:latest

# If installed locally in project:
npx sieveai check

Configuration

Create a .sieveai.config.json file in your project root:

{
  "model": "ollama:gemma3:latest",
  "temperature": 0.7,
  "maxTokens": 2048,
  "minConfidence": 78,
  "enableCache": true,
  "parallel": true,
  "maxConcurrency": 3,
  "enableSecretScanning": true,
  "format": "tui",
  "verbose": false,
  "showRationale": true
}

Configuration Options

| Option | Type | Default | Description | |--------|------|---------|-------------| | model | string | "ollama:gemma3:latest" | LLM model (provider:model:tag) | | temperature | number | 0.7 | AI creativity (0.0-1.0) | | maxTokens | number | 2048 | Maximum response length | | minConfidence | number | 78 | Minimum confidence for findings (0-100) | | enableCache | boolean | true | Enable result caching | | parallel | boolean | true | Run agents in parallel | | maxConcurrency | number | 3 | Max parallel agent execution | | enableSecretScanning | boolean | true | Scan for leaked secrets | | format | string | "tui" | Output format: "tui", "json", "text" | | verbose | boolean | false | Enable detailed logging | | showRationale | boolean | true | Show reasoning for findings |

Usage Examples

Example 1: C++ Code Review

// ProcessingUnit.cpp
class ProcessingUnit {
public:
    void execute(std::vector<int> data) {  // ❌ Missing const, pass by value
        if (data.size() > 0) {             // ❌ Use !data.empty()
            process(data[0]);
        }
    }
private:
    std::string secret_key = "hardcoded_key"; // ❌ Security issue
};

SieveAi Analysis:

● HIGH | PERFORMANCE | ProcessingUnit.cpp:3
  Function parameter should be const reference
  Fix: Change to `const std::vector<int>& data` for better performance

● MEDIUM | STYLE | ProcessingUnit.cpp:4  
  Use container.empty() instead of size() comparison
  Fix: Replace `data.size() > 0` with `!data.empty()`

● CRITICAL | SECURITY | ProcessingUnit.cpp:8
  Hardcoded secret detected in source code
  Fix: Move sensitive data to environment variables or config files

Example 2: JavaScript Security Analysis

// server.js
app.get('/user/:id', (req, res) => {
    const query = `SELECT * FROM users WHERE id = ${req.params.id}`;  // ❌ SQL injection
    db.query(query, (err, results) => {
        res.json(results);  // ❌ No error handling
    });
});

SieveAi Analysis:

● CRITICAL | SECURITY | server.js:2
  SQL injection vulnerability detected
  Fix: Use parameterized queries: `SELECT * FROM users WHERE id = ?`

● HIGH | BUG | server.js:3
  Unhandled database error could crash application  
  Fix: Add proper error handling for database operations

Advanced Features

Custom Analysis Depth

# High confidence only (fewer, higher-quality findings)
sieveai check --min-confidence 90

# Lower confidence (more findings, some potential false positives)  
sieveai check --min-confidence 50

# Performance-focused analysis
sieveai check --model ollama:deepseek-coder:latest --temperature 0.3

Git Integration

# Review specific commit range
sieveai check --base main --target feature-branch

# Review staged changes only
git add . && sieveai check --staged

# Review specific files
git diff path/to/file.cpp | sieveai check

Output Formats

# Beautiful TUI output (default)
sieveai check --format tui

# JSON for automation
sieveai check --format json | jq '.findings[] | select(.severity == "critical")'

# Plain text for logging
sieveai check --format text >> code-review.log

Development

Building from Source

# Development setup
git clone https://github.com/Jayesh-Dev21/SieveAi.git
cd SieveAi
npm install

# Development build with watch
npm run dev

# Production build
npm run build

# Run tests
npm test

Available Scripts

# Core scripts
npm run build          # Build TypeScript to dist/
npm run dev           # Watch mode for development
npm test             # Run test suite
npm run lint         # Lint code with ESLint
npm run format       # Format code with Prettier

# Quick usage
npm run check         # Run code review
npm run check:verbose # Run with verbose output
npm run check:json    # Get JSON output
npm run setup         # One-command setup

Project Structure

src/
├── agents/          # AI analysis agents (bug, security, style)
├── cache/           # Result caching system
├── cli/             # Command-line interface
├── config/          # Configuration management
├── git/             # Git integration and diff parsing
├── llm/             # LLM client and prompts
├── orchestrator/    # Main review coordination
├── reporters/       # Output formatting (TUI, JSON, text)
├── static/          # Static analyzers (syntax, secrets, whitespace)
├── types/           # TypeScript type definitions
└── utils/           # Shared utilities

docs/               # Documentation
examples/           # Usage examples
web/                # Web interface (Vite + React)
dist/               # Compiled output

Troubleshooting

Common Issues

LLM Not Available

# Check if Ollama is running
ollama list

# Restart Ollama service
systemctl restart ollama  # Linux
brew services restart ollama  # macOS

Model Not Found

# Pull the model explicitly
ollama pull gemma3:latest

# Check available models
ollama list

Slow Analysis

# Reduce model size
sieveai check --model ollama:gemma3:2b  

# Disable caching temporarily
sieveai check --no-cache

# Reduce concurrency
sieveai check --max-concurrency 1

High Memory Usage

# Use smaller model
ollama pull gemma3:2b

# Reduce token limit
echo '{"maxTokens": 1024}' > .sieveai.config.json

🤝 Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

Areas for Contribution

• New Language Support: Add syntax checkers for additional languages
• AI Agents: Enhance existing agents or create specialized ones
• LLM Providers: Add support for OpenAI, Anthropic, etc.
• Static Analyzers: Improve existing or add new static analysis tools
• Documentation: Help improve documentation and examples

License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Ollama for local LLM inference
• Google Gemma for the base AI model
• All contributors and the open source community

Made with ❤️ for better code quality