toxo-docs

🚀 TOXO Quickstart — From Zero to AI Expert in 2 Minutes

Transform any LLM into a CALM (Context Augmented Language Model) faster than making coffee. No GPUs, no fine-tuning, no $100K bills.

⚡ Installation (30 seconds)

# Core library
pip install toxo

# Install by use case (recommended)
pip install "toxo[openai]"        # OpenAI provider support
pip install "toxo[claude]"        # Anthropic provider support
pip install "toxo[rag]"           # Vector store + retrieval
pip install "toxo[vision]"        # Optional local document/image tooling
pip install "toxo[auth]"          # JWT + bcrypt
pip install "toxo[distributed]"   # Redis-backed features
pip install "toxo[enterprise]"    # Monitoring/telemetry (optional)
pip install "toxo[full]"          # Everything (largest install)

Note: Multimodal via Gemini works out of the box in direct mode (bytes sent straight to Gemini). Local PDF rendering/OCR (toxo[vision] + Poppler) is optional.

🎯 Your First Smart Layer (60 seconds)

from toxo import ToxoLayer

# Load your trained domain expert
layer = ToxoLayer.load("financial_advisor.toxo")

# Connect to ANY LLM (user specifies model)
layer.setup_api_key(
    api_key="YOUR_GEMINI_API_KEY",
    model="gemini-2.5-flash-lite",  # Recommended default
    provider="gemini"
)

# Get expert-level responses instantly
response = layer.query("What's the best investment strategy for 2024?")
print(response)
# 🎯 Output: Professional financial analysis with risk assessment

🔥 Multi-Provider Magic

Switch LLMs Without Retraining

# Same smart layer, different LLMs!

# Use Google's Gemini (recommended)
layer.setup_api_key("key", "gemini-2.5-flash-lite", "gemini")

# Switch to OpenAI GPT-4 / GPT-4o
layer.setup_api_key("key", "gpt-4o", "openai")

# Try Anthropic Claude  
layer.setup_api_key("key", "claude-3.5-sonnet", "claude")

# Same expertise, any provider! 🚀

💡 Pro Tips for Viral Success

Context-Aware Queries

# Smart layers understand context
answer = layer.query(
    "Analyze this market trend",
    context={
        "user_role": "portfolio_manager", 
        "risk_level": "moderate",
        "timeframe": "6_months"
    }
)
# Gets personalized, role-specific analysis! 🎯

Production-Ready Async

import asyncio

async def main():
    legal_expert = ToxoLayer.load("legal_advisor.toxo")
    legal_expert.setup_api_key("key", "gpt-4", "openai")
    
    # Lightning-fast parallel processing
    result = await legal_expert.query_async(
        "Flag risky clauses in this contract"
    )
    print(result)

asyncio.run(main())

🏆 Success Stories

“Replaced our $50K fine-tuning project with a $50 TOXO layer. Same performance, 1000x cheaper!” — Senior ML Engineer

“Switched from GPT to Claude in 5 minutes without losing domain expertise.” — AI Startup Founder

🧪 Testing TOXO locally

Run the end-to-end capability runner from the toxo Python package repo:

# Fast (no network)
python3 testing/toxo_capability_test.py --layer examples/minimal_test_layer.toxo

# Full (live Gemini + multimodal + evidence tests)
python3 testing/toxo_capability_test.py \
  --layer examples/rich_test_layer.toxo \
  --live --multimodal --evidence

To create a richer local test layer:

python3 testing/create_rich_test_layer.py

⚠️ Important Notes

• Model names required: No hidden defaults. YOU control which model to use.
• API keys needed: Get keys from your chosen provider (Google Gemini, OpenAI, Anthropic).
• No GPUs required: Works entirely through LLM APIs.

🚀 What’s Next?

• 📖 API Reference — Complete method documentation
• 📄 Document Conversion — DOCX, HTML, TXT → PDF for Gemini
• ☁️ Cloud Deployment — Deploy to Google Cloud Run
• 🔧 Provider Setup — Multi-provider configuration
• 📋 FAQ — Common questions answered

Ready to revolutionize your AI workflow? You’re just one pip install toxo away! 🎉