Integrating Groq, DeepInfra, and Nebius in One Edge Function — 3-Provider AI Routing

Integrating Groq, DeepInfra, and Nebius in One Edge Function Why Three Providers? Each AI provider has different strengths: Provider Strength Model Groq Extremely fast (custom Inference Engine) Llama 3.3 70B DeepInfra 400+ models, lowest cost Llama 3.3 70B Turbo Nebius High quality prose generation Llama 3.1 70B Rather than three separate Edge Functions, I integrated all three into a single ai-hub EF using an action dispatch pattern. Architecture: PROVIDER_CONFIGS Map // supabase/functions/ai-hu