toxo-docs

🌟 Multi-Provider LLM Support — Your Choice, Maximum Flexibility

TOXO smart layers work with ANY LLM provider. Train once, deploy everywhere. Switch models in seconds without losing expertise.

🤖 Google Gemini (Recommended Default)

Why Gemini?

• ⚡ Fastest responses - Optimized for speed
• 💰 Cost-effective - Great price/performance ratio
• 🧠 Latest AI - Google’s cutting-edge models

Setup

layer.setup_api_key(
    api_key="YOUR_GEMINI_KEY",
    model="gemini-2.5-flash-lite",  # Recommended default
    provider="gemini"
)

Popular Models

• gemini-2.5-flash-lite - Default: fast + cost-effective
• gemini-1.5-pro - Most capable (reasoning)
• gemini-1.5-flash - Balanced (speed + cost)

Get API Key

1. Visit Google AI Studio
2. Create project & enable Gemini API
3. Generate API key

🧠 OpenAI GPT (Most Popular)

Why OpenAI?

• 🏆 Industry leader - Proven performance
• 📚 Best documentation - Extensive resources
• 🔧 Developer-friendly - Great tooling

Setup

layer.setup_api_key(
    api_key="YOUR_OPENAI_KEY",
    model="gpt-5.4-thinking",  # Latest high-reasoning model
    provider="openai"
)

Popular Models

• gpt-5.4-thinking - Most factual, advanced reasoning
• gpt-5.4-pro - Production-ready, balanced cost/quality
• gpt-5.3-codex - Coding-focused agent

Get API Key

1. Visit OpenAI Platform
2. Create account & add billing
3. Generate API key

🎭 Anthropic Claude (Analysis Expert)

Why Claude?

• 📊 Best at analysis - Excels at complex reasoning
• 🔍 Detail-oriented - Thorough and accurate
• 🛡️ Safety-focused - Built-in safety measures

Setup

layer.setup_api_key(
    api_key="YOUR_ANTHROPIC_KEY",
    model="claude-sonnet-4-6",  # Analysis champion
    provider="claude"
)

Popular Models

• claude-opus-4-6 - Most capable (reasoning + coding)
• claude-sonnet-4-6 - Best balance of speed and quality
• claude-haiku-4.5 - Lightweight, fast tasks

Get API Key

1. Visit Anthropic Console
2. Create account & add credits
3. Generate API key

🔄 Switch Providers Instantly

Same Layer, Different LLMs

# Load your expert once
financial_expert = ToxoLayer.load("finance_advisor.toxo")

# Try Gemini for speed (recommended)
financial_expert.setup_api_key("key", "gemini-2.5-flash-lite", "gemini")
response1 = financial_expert.query("Market analysis")

# Switch to GPT-4 / GPT-4o for reasoning  
financial_expert.setup_api_key("key", "gpt-4o", "openai")
response2 = financial_expert.query("Complex financial modeling")

# Use Claude for detailed analysis
financial_expert.setup_api_key("key", "claude-3.5-sonnet", "claude")  
response3 = financial_expert.query("Risk assessment report")

🎯 Choosing the Right Provider

For Speed & Cost

• Winner: Gemini (gemini-2.5-flash-lite)
• Use case: High-volume, real-time applications

For Reasoning & Logic

• Winner: OpenAI (gpt-4o)
• Use case: Complex problem-solving, coding

For Analysis & Detail

• Winner: Claude (claude-3.5-sonnet)
• Use case: Document analysis, research

For Multimodal Tasks

• Winner: OpenAI (gpt-4o)
• Use case: Image + text processing

⚠️ Important Notes

• Model names required: Always specify the exact model
• No hidden defaults: YOU choose which model to use
• API keys needed: Get from your chosen provider
• Rate limits apply: Check provider documentation

🚀 Pro Tips

Cost Optimization

# Use faster models for simple queries
layer.setup_api_key("key", "gemini-1.5-flash", "gemini")

# Switch to powerful models for complex tasks
layer.setup_api_key("key", "gpt-4o", "openai")

Provider Fallback

try:
    layer.setup_api_key("primary_key", "gpt-4o", "openai")
    response = layer.query("Complex query")
except Exception:
    # Fallback to alternative provider
    layer.setup_api_key("backup_key", "claude-3.5-sonnet", "claude")
    response = layer.query("Complex query")

🌟 Future Providers

Coming soon:

• 🦙 Local Models (Ollama, LM Studio)
• 🏢 Enterprise APIs (Azure OpenAI, AWS Bedrock)
• 🔧 Custom Endpoints (Your own models)

The beauty of TOXO: Your smart layers will work with ALL future providers! 🚀