THE.MAINREPO.01

AI Provider Integrations

This document explains how to configure and use multiple AI providers with Ai-bilder.

Supported Providers

Ai-bilder supports the following AI providers:

• OpenAI (GPT-3.5, GPT-4, and newer models)
• Anthropic (Claude models)
• Azure OpenAI (Microsoft-hosted OpenAI models)

Configuration

Environment Variables

Configure your AI provider using environment variables in your .env file:

# Select your primary AI provider
AI_PROVIDER=openai  # Options: openai, anthropic, azure-openai

# OpenAI Configuration
OPENAI_API_KEY=your-openai-api-key-here

# Anthropic Configuration
ANTHROPIC_API_KEY=your-anthropic-api-key-here

# Azure OpenAI Configuration
AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com/
AZURE_OPENAI_KEY=your-azure-openai-key-here

# Rate Limiting
RATE_LIMIT_QPS=10  # Queries per second (adjust based on your API tier)

# Caching
CACHE_TTL_SECONDS=60  # Cache responses for 60 seconds

Provider-Specific Setup

OpenAI

1. Create an account at platform.openai.com
2. Generate an API key from your account settings
3. Set AI_PROVIDER=openai and OPENAI_API_KEY in your .env

Anthropic (Claude)

1. Sign up at console.anthropic.com
2. Create an API key
3. Set AI_PROVIDER=anthropic and ANTHROPIC_API_KEY in your .env

Azure OpenAI

1. Set up an Azure OpenAI resource in the Azure Portal
2. Deploy a model (e.g., GPT-4)
3. Get your endpoint and API key
4. Set AI_PROVIDER=azure-openai, AZURE_OPENAI_ENDPOINT, and AZURE_OPENAI_KEY in your .env

Rate Limiting

Each AI provider has different rate limits based on your subscription tier:

• OpenAI: Varies by tier (free, pay-as-you-go, enterprise)
• Anthropic: Varies by plan
• Azure OpenAI: Based on your Azure subscription and deployment

Configure RATE_LIMIT_QPS (Queries Per Second) to match your provider’s limits:

RATE_LIMIT_QPS=10  # Conservative default
# For higher tiers, you might use:
# RATE_LIMIT_QPS=50  # Pro tier
# RATE_LIMIT_QPS=100 # Enterprise tier

Best Practices for Rate Limiting

• Start with a conservative limit and increase gradually
• Monitor your API usage in your provider’s dashboard
• Implement exponential backoff for failed requests
• Use caching to reduce API calls

Caching

Ai-bilder includes built-in response caching to reduce costs and improve performance:

CACHE_TTL_SECONDS=60  # Cache responses for 60 seconds

Caching Strategies

• Short TTL (30-60s): For dynamic, frequently-changing content
• Medium TTL (5-15 min): For semi-static content
• Long TTL (1+ hour): For static or rarely-changing content

When to Use Caching

✅ Good use cases:

• Repeated queries with identical inputs
• Static content generation
• Template-based responses

❌ Avoid caching for:

• Real-time data requests
• User-specific personalized content
• Time-sensitive information

Retry Logic

Ai-bilder implements automatic retry logic for transient failures:

• Retry attempts: 3 by default
• Backoff strategy: Exponential with jitter
• Retry conditions: Network errors, rate limits, temporary server errors

Error Handling

Different error types are handled differently:

• Rate Limit Errors (429): Automatically retried with backoff
• Authentication Errors (401): Not retried - check your API keys
• Server Errors (500, 502, 503): Retried up to 3 times
• Client Errors (400): Not retried - check your request format

Bring Your Own Model (BYOM)

Ai-bilder supports using your own model endpoints. This is useful for:

• Custom fine-tuned models
• Self-hosted models
• Alternative providers not listed above

Configuration

Set environment variables for your custom endpoint:

AI_PROVIDER=custom
CUSTOM_AI_ENDPOINT=https://your-model-endpoint.com/v1
CUSTOM_AI_KEY=your-api-key

Monitoring and Debugging

API Usage Tracking

Monitor your API usage through:

• Provider dashboards (OpenAI, Anthropic, Azure)
• Application logs
• Cost tracking tools

Common Issues

Issue: Authentication errors

• Solution: Verify your API keys are correct and active

Issue: Rate limit exceeded

• Solution: Reduce RATE_LIMIT_QPS or upgrade your API tier

Issue: Slow responses

• Solution: Implement caching or use faster models

Issue: High costs

• Solution: Use caching, optimize prompts, or switch to cheaper models

Security Best Practices

1. Never commit API keys to version control
   • Always use .env files (included in .gitignore)
   • Use environment variables in production
2. Rotate keys regularly
   • Set up a key rotation schedule
   • Revoke old keys after rotation
3. Use least privilege
   • Only grant necessary permissions
   • Use separate keys for different environments
4. Monitor for anomalies
   • Set up alerts for unusual usage patterns
   • Review API logs regularly
5. Secure your environment
   • Use HTTPS for all API communications
   • Implement proper authentication in your application
   • Keep dependencies up to date

Provider Comparison

Feature	OpenAI	Anthropic	Azure OpenAI
Model Variety	✅ Extensive	✅ Claude series	✅ OpenAI models
Pricing	💰 Pay-per-use	💰 Pay-per-use	💰 Azure pricing
Rate Limits	Tier-based	Tier-based	Custom
Enterprise Support	✅ Yes	✅ Yes	✅ Yes
Data Privacy	Standard	Enhanced	Azure compliance
Custom Models	✅ Fine-tuning	Limited	✅ Fine-tuning

Additional Resources

• OpenAI API Documentation
• Anthropic API Documentation
• Azure OpenAI Documentation

Support

For issues specific to AI provider integration:

1. Check provider status pages
2. Review API documentation
3. Check application logs
4. Open an issue in this repository with relevant details

This site is open source. Improve this page.