🌐 WEVIA V5+ Multi-Server Dispatch DOCTRINE 314 WIRED

MSD: routing intelligent S204/S95/S151/Sovereign/Blade · Auto-décharge S204 si load > 20 · 7 routes possibles · Smart task-based dispatch

🛣 Routes intelligentes (task → server)

/(deepseek|chatgpt|claude.ai|gemini.web)/

→ blade_web

Web chats Chrome user cookies (free unlimited)

→ sovereign:4000

LLM API tasks (free APIs cascade 21 providers)

/(playwright|browser|screenshot|chrome|video)/

→ S95

Browser farm + sentinel-brain

/(track|email|open|click|pmta)/

→ S151 OVH

Tracking emails/opens/clicks

load > 20 || disk > 92%

→ Auto-decharge

Si S204 surchargé, route ailleurs auto

prefer="local"

→ S204 local

Force exécution locale (override)

sovereign HTTP 503 / busy

→ Fallback blade_web

Doctrine 333: cascade épuisée → cookies user (free)

🧪 Test bench MSD (curl exemples)

📤 Test sovereign route:
curl -X POST /api/multi-server-dispatch.php -d '{"task":"chat","prefer":"sovereign","params":{"messages":[{"role":"user","content":"hi"}]}}'

📤 Test S95 route:
curl -X POST /api/multi-server-dispatch.php -d '{"task":"playwright_screenshot","prefer":"auto"}'

📤 Test S151 route:
curl -X POST /api/multi-server-dispatch.php -d '{"task":"track_email_open","prefer":"auto"}'

📤 Test blade web cookie:
curl -X POST /api/multi-server-dispatch.php -d '{"task":"deepseek_chat","prefer":"blade_web"}'

📊 MSD Stats (24h)

Calls per route loading...