SetupAgent AI

An intelligent environment setup agent specializing in Node.js, Python, and Java projects. Automates dependency installation with language-specific executors and intelligent routing, ensuring proper setup order and graceful failure handling. Perfect for polyglot projects, CI/CD pipelines, and rapid onboarding.

Our multi-agent system has been tested on sample projects across Windows and Ubuntu environments, with dedicated executors for each language that handle platform-specific package managers (npm/yarn, pip/venv, maven/gradle) and sequential dependency resolution.

Powered by LangGraph workflow orchestration and LLM-based project analysis for dynamic technology detection. Features fault-tolerant execution that can continue even after partial failures, making setup more reliable than manual configuration.

Supported Platforms

• Windows 10/11 - Requires Administrator privileges
• Ubuntu 18.04+ - Debian-based Linux with sudo access

Getting Started

Install dependencies

Windows (Must Run as Administrator):

# Option 1: Right-click START_WINDOWS_ONE_CLICK.bat → Run as Administrator
# Option 2: Open PowerShell as Administrator and run:
.\START_WINDOWS_ONE_CLICK.bat

Ubuntu:

git clone https://github.com/your-repo/setup_agent
cd setup_agent
sudo apt-get update
sudo apt-get install -y python3-venv python3-pip
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
export OPENAI_API_KEY=sk-...      # or ANTHROPIC_API_KEY=sk-ant-api...

Running the Agent

python -m src.cli https://github.com/user/repo     # GitHub repository
python -m src.cli C:\path\to\project              # Local project
python -m src.cli .                               # Current directory

Performance

Tested on 40 projects: 95% success rate (38/40) with 3.5 min average setup time.

• Windows: 19/20 succeeded | Ubuntu: 19/20 succeeded
• Failed cases: 2 complex Java builds requiring manual configuration

Supported Platforms

• Windows 10/11 (Administrator required)
• Ubuntu 18.04+ / Debian-based Linux

Tested Project Types

• ✅ Python - Flask, Django, FastAPI
• ✅ Node.js - Express, React, Vue
• ✅ Java - Spring Boot, Maven/Gradle
• ✅ Multi-language - Full-stack applications

Setup Time Distribution

< 2 min:  ████████████ 30%
2-5 min:  ████████████████████ 50%
5-10 min: ████████ 20%

Agent Architecture

Orchestrator → Scanner → Analyzer → Planner → Language Executors (Loop) → Verifier → Reporter
     ↓                                      ↑                    ↓
[Metrics Init]                             ←─────────────────────

Execution Flow: Planner sequentially calls each language executor (Java → Python → Node.js), with automatic retry and failure isolation. Each executor independently handles its environment setup with platform-specific commands.

Multi-Agent System:

• Orchestrator: Initializes metrics & error tracking, coordinates workflow
• Scanner: Detects Node.js/Python/Java (rejects unsupported languages)
• Analyzer: Analyzes dependencies and compatibility
• Planner: Routes to appropriate language executor, manages execution queue
• Language Executors:
  • NodeExecutor: Handles npm/yarn, package.json dependencies
  • PythonExecutor: Manages pip/venv, requirements.txt/Pipfile
  • JavaExecutor: Configures JDK, Maven/Gradle builds
• Verifier: Validates all language environments
• Reporter: Generates setup report + exports metrics/errors

Data Tracking:

• Metrics: Real-time performance tracking per language (success rate, duration, commands)
• Error Classification: Categorized errors (NETWORK_ERROR, PERMISSION_DENIED, TIMEOUT, etc.)
• Output Files:
  • reports/setup_[project]_[timestamp].md - Human-readable report
  • metrics/setup_metrics_[timestamp].json - Performance data
  • reports/errors_[project]_[timestamp].json - Error details (if any)