Senly runs all AI calls through a ladder of six (free) providers in this order:
1. Gemini (Google AI)
2. Groq (fast via LPUs)
3. Cerebras (Wafer-Scale, ultra-fast)
4. Cloudflare Workers AI
5. GitHub Models
6. OpenRouter
If the first one fails, Senly automatically tries the second. Result: you almost never experience AI downtime, even if one provider is down.