AI Providers in GateFlow

GateFlow supports multiple AI providers with unified sustainability optimization across all platforms.

Supported Providers

General AI Providers

• OpenAI - GPT models, DALL·E, Whisper
• Anthropic - Claude models
• Google - Gemini models
• Mistral - Mistral and Mixtral models
• Cohere - Efficient chat, embeddings, and rerank
• ElevenLabs - Text-to-speech and voice synthesis

Sustainability Features by Provider

Provider-Specific Sustainability Features

Provider	Unique Sustainability Features
Cohere	Most carbon-efficient models, specialized embeddings, rerank optimization
ElevenLabs	Low-carbon TTS, batch processing, time-shifted audio generation
Anthropic	High efficiency per token, excellent quality-to-carbon ratio
Mistral	Open-source efficiency, European data centers
OpenAI	Broad model range, carbon offset programs
Google	Data center efficiency, renewable energy commitments

Multi-Provider Sustainability Strategies

Optimal Provider Selection Guide

Use Case	Recommended Primary	Fallback Option
Chat Applications	Anthropic Claude	OpenAI GPT-5.2
Text-to-Speech	OpenAI TTS	Google TTS
Embeddings	OpenAI text-embedding-3	Mistral embeddings
RAG Systems	OpenAI embeddings	Mistral embeddings
Complex Analysis	Anthropic Claude	OpenAI GPT-5.2

Configuration Example

from openai import OpenAI

client = OpenAI(
    base_url="https://api.gateflow.ai/v1",
    api_key="gw_prod_your_key_here"
)

# Let GateFlow choose the most sustainable provider automatically
response = client.chat.completions.create(
    model="auto",  # Auto-select across all providers
    routing_mode="sustain_optimized",
    minimum_quality_score=85,
    messages=[{"role": "user", "content": "Analyze this sustainably"}]
)

print(f"Selected provider: {response.provider}")
print(f"Model used: {response.model}")
print(f"Carbon saved: {response.sustainability.carbon_saved_gco2e} gCO₂e")

Provider Setup Guide

Adding Providers

1. Navigate to Dashboard: Go to Settings → Providers
2. Add Provider: Click "Add Provider" and select from the list
3. Enter API Key: Paste your provider API key
4. Configure Settings: Set region preferences and sustainability options
5. Save: The provider is now available for routing

Provider-Specific Settings

Each provider supports custom configuration:

• Region Preferences: Prioritize data centers in low-carbon regions
• Model Preferences: Set default models for different task types
• Quality Thresholds: Define minimum quality scores for Sustain Mode
• Carbon Budgets: Set maximum emissions per request type

Advanced Multi-Provider Routing

Custom Routing Rules

# Configure in Dashboard: Routing → Custom Rules
rules = [
    {
        "task_type": "text_to_speech",
        "primary_provider": "elevenlabs",
        "fallback_providers": ["openai", "google"],
        "routing_mode": "sustain_optimized"
    },
    {
        "task_type": "semantic_search",
        "primary_provider": "cohere",
        "fallback_providers": ["openai", "mistral"],
        "routing_mode": "sustain_optimized",
        "minimum_quality_score": 80
    }
]

Provider Fallback Chains

Monitoring Provider Performance

Sustainability Dashboard

Track carbon efficiency across providers:

# Get provider sustainability metrics
provider_metrics = client.sustainability.providers.list()

for provider in provider_metrics.providers:
    print(f"{provider.name}:")
    print(f"  Avg carbon/token: {provider.avg_carbon_per_token_gco2e} gCO₂e")
    print(f"  Requests this month: {provider.request_count}")
    print(f"  Total carbon saved: {provider.total_carbon_saved_gco2e} gCO₂e")
    print(f"  Efficiency score: {provider.efficiency_score}/100")

Provider Efficiency Reports

Generate detailed reports in the Dashboard:

1. Navigate to Analytics → Provider Efficiency
2. Select Time Range: Last 7 days, 30 days, or custom
3. View Metrics: Carbon efficiency, cost, quality, latency
4. Export Data: CSV or PDF for ESG reporting

Best Practices for Multi-Provider Sustainability

Optimization Strategies

1. Diversify Providers: Use multiple providers for resilience and optimization
2. Set Quality Thresholds: Balance carbon savings with quality requirements
3. Monitor Grid Conditions: Adjust routing based on real-time carbon intensity
4. Use Time-Shifting: Defer non-urgent requests to low-carbon periods
5. Enable Caching: Reduce duplicate computations across providers
6. Regular Review: Update provider preferences as new models become available

Common Pitfalls to Avoid

❌ Over-optimizing: Don't sacrifice critical quality for minimal carbon savings ❌ Single Provider Dependency: Avoid relying on one provider for all workloads ❌ Ignoring Regional Differences: Carbon intensity varies by data center location ❌ Static Configuration: Regularly update settings as providers improve efficiency

Next Steps

• Explore Cohere Integration - Most carbon-efficient provider
• Try ElevenLabs TTS - Low-carbon voice synthesis
• Configure Sustain Mode - Automatic provider optimization
• View Provider Analytics - Monitor provider performance