Environment Configuration

Environment File Structure

LiteAgent uses environment files to manage API keys, configuration settings, and runtime parameters. The main configuration is stored in collector/.env.

Setting Up the Environment File

Copy Example File

cd agent-collector
cp collector/.env.example collector/.env

Edit Configuration

# Use your preferred editor
nano collector/.env
# or
code collector/.env
# or
vim collector/.env

Verify Configuration

# Check that required keys are set
python -c "
import os
from dotenv import load_dotenv
load_dotenv('collector/.env')

required_keys = ['OPENAI_API_KEY']
for key in required_keys:
    if os.getenv(key):
        print(f'✓ {key} is set')
    else:
        print(f'✗ {key} is missing')
"

API Keys Configuration

Primary LLM Providers

OpenAI
Anthropic
Google

Required for: BrowserUse, WebArena, VisualWebArena

# OpenAI API Key
OPENAI_API_KEY=sk-proj-...

# Optional: Organization ID
OPENAI_ORG_ID=org-...

# Optional: Model preferences
OPENAI_DEFAULT_MODEL=gpt-4-vision-preview
OPENAI_TEMPERATURE=0.7
OPENAI_MAX_TOKENS=4000

Getting Your Key:

Visit platform.openai.com
Create a new API key
Copy the key (starts with sk-proj- or sk-)

Pricing: ~$0.01-0.06 per 1K tokens depending on model

Agent-Specific APIs

MultiOn
DoBrowser
Skyvern

Required for: MultiOn agent

# MultiOn API Key
MULTION_API_KEY=...

# Optional: MultiOn settings
MULTION_TIMEOUT=60
MULTION_MAX_RETRIES=3

Getting Your Key:

Visit multion.ai
Sign up for an account
Generate API key from dashboard

Application Settings

Core Configuration

# Application settings
APP_NAME=LiteAgent
VERSION=1.0.0
DEBUG=false
LOG_LEVEL=INFO

# Default timeouts (seconds)
DEFAULT_TIMEOUT=180
PAGE_LOAD_TIMEOUT=30
ELEMENT_WAIT_TIMEOUT=10

# Retry settings
MAX_RETRIES=3
RETRY_DELAY=2

Browser Configuration

# Browser settings
HEADLESS=true
BROWSER_WIDTH=1920
BROWSER_HEIGHT=1080

# Chrome arguments (comma-separated)
BROWSER_ARGS=--no-sandbox,--disable-dev-shm-usage,--disable-gpu

# Browser binary path (optional)
CHROME_EXECUTABLE_PATH=/usr/bin/google-chrome

# User agent override (optional)
USER_AGENT=Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36

Storage Configuration

# Data directories
DATA_DIR=./data
OUTPUT_DIR=./data/db
PROMPT_DIR=./data/prompts
LOG_DIR=./collector/logs
BROWSER_DATA_DIR=./data/browser_data

# Database settings
DB_MAX_CONNECTIONS=10
DB_TIMEOUT=30

# Video recording
ENABLE_VIDEO_RECORDING=true
VIDEO_QUALITY=high

Network Configuration

# Proxy settings (optional)
HTTP_PROXY=http://proxy.example.com:8080
HTTPS_PROXY=http://proxy.example.com:8080
NO_PROXY=localhost,127.0.0.1

# Request timeouts
HTTP_TIMEOUT=30
DOWNLOAD_TIMEOUT=300

# Rate limiting
REQUESTS_PER_MINUTE=60
BURST_LIMIT=10

Environment-Specific Configurations

Development Environment

Create collector/.env.dev:

# Development settings
DEBUG=true
LOG_LEVEL=DEBUG
HEADLESS=false

# Faster timeouts for development
DEFAULT_TIMEOUT=60
MAX_RETRIES=1

# Local webhook for testing
SKYVERN_WEBHOOK_URL=http://localhost:8080

# Verbose logging
VERBOSE_LOGGING=true
SAVE_SCREENSHOTS=true

Production Environment

Create collector/.env.prod:

# Production settings
DEBUG=false
LOG_LEVEL=WARNING
HEADLESS=true

# Production timeouts
DEFAULT_TIMEOUT=300
MAX_RETRIES=5

# Production webhook
SKYVERN_WEBHOOK_URL=https://your-domain.com/webhook

# Optimized for performance
ENABLE_VIDEO_RECORDING=false
SAVE_SCREENSHOTS=false

Testing Environment

Create collector/.env.test:

# Test settings
DEBUG=true
LOG_LEVEL=INFO
HEADLESS=true

# Fast tests
DEFAULT_TIMEOUT=30
MAX_RETRIES=1

# Test data isolation
OUTPUT_DIR=./data/test_db
LOG_DIR=./collector/test_logs

# Mock APIs for testing
USE_MOCK_APIS=true

Loading Environment Variables

Using with Python

from dotenv import load_dotenv
import os

# Load default environment
load_dotenv('collector/.env')

# Load environment-specific config
env = os.getenv('ENV', 'dev')
load_dotenv(f'collector/.env.{env}')

# Access variables
openai_key = os.getenv('OPENAI_API_KEY')
timeout = int(os.getenv('DEFAULT_TIMEOUT', 180))

Using with Docker

# docker-compose.yml
services:
  browseruse:
    env_file:
      - collector/.env
      - collector/.env.${ENV:-dev}
    environment:
      - ENV=${ENV:-dev}

Using with Shell Scripts

#!/bin/bash
# load_env.sh

# Load environment file
if [ -f collector/.env ]; then
    export $(grep -v '^#' collector/.env | xargs)
fi

# Load environment-specific file
ENV=${ENV:-dev}
if [ -f "collector/.env.$ENV" ]; then
    export $(grep -v '^#' "collector/.env.$ENV" | xargs)
fi

echo "Environment loaded: $ENV"

Security Best Practices

API Key Security

Never commit API keys to version control!

# Add .env files to .gitignore
echo "collector/.env*" >> .gitignore
echo "!collector/.env.example" >> .gitignore

# Verify .env files are ignored
git status  # Should not show .env files

Key Rotation

# Script to rotate API keys
#!/bin/bash
# rotate_keys.sh

echo "Rotating API keys..."

# Backup current config
cp collector/.env collector/.env.backup.$(date +%Y%m%d)

# Update keys (replace with actual new keys)
sed -i 's/OPENAI_API_KEY=.*/OPENAI_API_KEY=sk-new-key/' collector/.env
sed -i 's/ANTHROPIC_API_KEY=.*/ANTHROPIC_API_KEY=sk-ant-new-key/' collector/.env

echo "Keys rotated. Test new configuration:"
python -c "
from dotenv import load_dotenv
import os
load_dotenv('collector/.env')
print('✓ New keys loaded')
"

Environment Validation

Create a validation script:

# scripts/validate_env.py
import os
from dotenv import load_dotenv

def validate_environment():
    load_dotenv('collector/.env')

    # Required keys
    required_keys = {
        'OPENAI_API_KEY': 'OpenAI API key',
        'ANTHROPIC_API_KEY': 'Anthropic API key (for Agent E)',
    }

    # Optional keys
    optional_keys = {
        'MULTION_API_KEY': 'MultiOn API key',
        'DOBROWSER_API_KEY': 'DoBrowser API key',
        'SKYVERN_API_KEY': 'Skyvern API key',
    }

    print("🔍 Environment Validation")
    print("=" * 30)

    # Check required keys
    missing_required = []
    for key, description in required_keys.items():
        value = os.getenv(key)
        if value:
            print(f"✅ {key}: {description}")
        else:
            print(f"❌ {key}: {description} (MISSING)")
            missing_required.append(key)

    # Check optional keys
    print("\nOptional configurations:")
    for key, description in optional_keys.items():
        value = os.getenv(key)
        if value:
            print(f"✅ {key}: {description}")
        else:
            print(f"⚠️  {key}: {description} (not set)")

    # Validate settings
    timeout = os.getenv('DEFAULT_TIMEOUT', '180')
    try:
        timeout_int = int(timeout)
        if timeout_int > 0:
            print(f"✅ DEFAULT_TIMEOUT: {timeout_int}s")
        else:
            print("❌ DEFAULT_TIMEOUT: Must be positive")
    except ValueError:
        print("❌ DEFAULT_TIMEOUT: Must be a number")

    # Summary
    if missing_required:
        print(f"\n❌ Missing {len(missing_required)} required keys")
        return False
    else:
        print("\n✅ All required keys configured")
        return True

if __name__ == "__main__":
    is_valid = validate_environment()
    exit(0 if is_valid else 1)

Run validation:

python scripts/validate_env.py

Configuration Templates

Complete Example

Here’s a complete .env file template:

# =============================================================================
# LiteAgent Configuration
# =============================================================================

# -----------------------------------------------------------------------------
# LLM Provider API Keys
# -----------------------------------------------------------------------------

# OpenAI (required for BrowserUse, WebArena, VisualWebArena)
OPENAI_API_KEY=sk-proj-your-key-here
OPENAI_ORG_ID=org-your-org-id
OPENAI_DEFAULT_MODEL=gpt-4-vision-preview

# Anthropic (required for Agent E)
ANTHROPIC_API_KEY=sk-ant-api03-your-key-here
ANTHROPIC_DEFAULT_MODEL=claude-3-5-sonnet-20241022

# Google (optional)
GOOGLE_API_KEY=AIza-your-key-here
GEMINI_API_KEY=AIza-your-key-here

# -----------------------------------------------------------------------------
# Agent-Specific API Keys
# -----------------------------------------------------------------------------

# MultiOn
MULTION_API_KEY=your-multion-key

# DoBrowser
DOBROWSER_API_KEY=your-dobrowser-key

# Skyvern
SKYVERN_API_KEY=your-skyvern-key
SKYVERN_WEBHOOK_URL=http://localhost:8080

# -----------------------------------------------------------------------------
# Application Settings
# -----------------------------------------------------------------------------

# Environment
ENV=dev
DEBUG=false
LOG_LEVEL=INFO

# Timeouts
DEFAULT_TIMEOUT=180
PAGE_LOAD_TIMEOUT=30
ELEMENT_WAIT_TIMEOUT=10

# Browser
HEADLESS=true
BROWSER_WIDTH=1920
BROWSER_HEIGHT=1080
BROWSER_ARGS=--no-sandbox,--disable-dev-shm-usage

# Storage
DATA_DIR=./data
OUTPUT_DIR=./data/db
PROMPT_DIR=./data/prompts
LOG_DIR=./collector/logs

# Features
ENABLE_VIDEO_RECORDING=true
SAVE_SCREENSHOTS=true
VERBOSE_LOGGING=false

Troubleshooting

API key not working

# Test OpenAI key
curl -H "Authorization: Bearer $OPENAI_API_KEY" \
     https://api.openai.com/v1/models | jq '.data[0].id'

# Test Anthropic key
curl -H "x-api-key: $ANTHROPIC_API_KEY" \
     https://api.anthropic.com/v1/messages \
     -d '{"model":"claude-3-haiku-20240307","max_tokens":10,"messages":[{"role":"user","content":"Hi"}]}'

Environment not loading

# Debug environment loading
import os
from dotenv import load_dotenv

print("Before loading:", os.getenv('OPENAI_API_KEY'))
result = load_dotenv('collector/.env')
print("Load result:", result)
print("After loading:", os.getenv('OPENAI_API_KEY'))

# Check file exists
import pathlib
env_file = pathlib.Path('collector/.env')
print("File exists:", env_file.exists())
print("File size:", env_file.stat().st_size if env_file.exists() else "N/A")

Permission errors

# Fix file permissions
chmod 600 collector/.env*  # Read/write for owner only
chown $USER:$USER collector/.env*

# Check permissions
ls -la collector/.env*

Next Steps

Running Tests

Start running tests with your configured environment

Agent Configuration

Configure individual agents

Docker Setup

Use environment with Docker deployment

Getting Started

Core Concepts

Setup & Configuration

Running Tests

Output & Analysis

Environment File Structure

Setting Up the Environment File

API Keys Configuration

Primary LLM Providers

Agent-Specific APIs

Application Settings

Core Configuration

Browser Configuration

Storage Configuration

Network Configuration

Environment-Specific Configurations

Development Environment

Production Environment

Testing Environment

Loading Environment Variables

Using with Python

Using with Docker

Using with Shell Scripts

Security Best Practices

API Key Security

Key Rotation

Environment Validation

Configuration Templates

Complete Example

Troubleshooting

Next Steps

Running Tests

Agent Configuration

Docker Setup

Getting Started

Core Concepts

Setup & Configuration

Running Tests

Output & Analysis

​Environment File Structure

​Setting Up the Environment File

​API Keys Configuration

​Primary LLM Providers

​Agent-Specific APIs

​Application Settings

​Core Configuration

​Browser Configuration

​Storage Configuration

​Network Configuration

​Environment-Specific Configurations

​Development Environment

​Production Environment

​Testing Environment

​Loading Environment Variables

​Using with Python

​Using with Docker

​Using with Shell Scripts

​Security Best Practices

​API Key Security

​Key Rotation

​Environment Validation

​Configuration Templates

​Complete Example

​Troubleshooting

​Next Steps

Running Tests

Agent Configuration

Docker Setup

Environment File Structure

Setting Up the Environment File

API Keys Configuration

Primary LLM Providers

Agent-Specific APIs

Application Settings

Core Configuration

Browser Configuration

Storage Configuration

Network Configuration

Environment-Specific Configurations

Development Environment

Production Environment

Testing Environment

Loading Environment Variables

Using with Python

Using with Docker

Using with Shell Scripts

Security Best Practices

API Key Security

Key Rotation

Environment Validation

Configuration Templates

Complete Example

Troubleshooting

Next Steps