HUBS 🧠 WEVIA πŸ€– AI πŸ‘₯ Agents πŸ“Š Monitor πŸ“§ Email πŸ“Š Office πŸ‘¨β€βš•οΈ Ethica πŸ“§ WEVADS ⚑ Blade πŸ›‘οΈ SΓ©cu ⚑ GPU πŸ” Keys ☁️ CF πŸ” Google 🌐 NC ⭐ ALL

🧠 AI Sovereign Hub

14 providers, 4 Ollama local, Qdrant RAG, cascade 0 EUR β€” souverainete totale

8
Tier 1
6
Tier 2
4
Ollama Local
16K+
Qdrant Vectors
0 EUR
Cout Total
59
Secrets
3
GPU Free

🚀 Tier 1 β€” Primary Cascade

Groq

llama-3.3-70b DEFAULT. 18 modeles disponibles. Latence ~200ms. Rate: 30 req/min

T1 DEFAULTFREE

HuggingFace Router

Qwen2.5-72B-Instruct. Inference API serverless. Fallback Groq

T1FREE

NVIDIA NIM

Nemotron-49B-Instruct. nvapi key active. GPU A100 cloud

T1FREE TIER

Cerebras

qwen-3-235b + 3 autres. Inference ultra-rapide ASIC. 4 modeles

T1FREE

SambaNova

DeepSeek-V3.1 + Llama-3.3. Custom silicon. Latence faible

T1FREE

Gemini 2.5 Flash

Google AI Studio. 1M tokens context. Multimodal (images+audio)

T1FREE TIER

Cloudflare Workers AI

Llama-3.1-8B + DeepSeek-R1-32B. GPU edge gratuit. 10K neurons/jour

T1GPU FREE

Mistral

Mistral-Large + Small. Paris-based. EU data sovereignty

T1FREE TIER

💻 Ollama Local (S204)

gemma4:e4b

Google Gemma 4 extended. 4B params. Port 11434 localhost

LOCAL

qwen3:4b

Alibaba Qwen 3. 4B params. Code + math + raisonnement

LOCAL

nomic-embed-text

Embeddings 768d pour Qdrant RAG. Semantic search

EMBEDDINGS

all-minilm

Sentence embeddings rapides. 384d. Classification + similarity

EMBEDDINGS

📦 RAG Qdrant

weval_skills

14,368 vecteurs β€” competences et patterns WEVIA. Base de connaissances principale

STABLE

wevia_learnings

1,390 vecteurs β€” apprentissages autonomes. +16 vec/heure via autolearn

AUTOLEARN

wevia_kb

386 vecteurs β€” knowledge base editoriale. Documentation technique

KB

wevia_memory

48 vecteurs β€” mΓ©moire conversationnelle. Context long-terme

MEMORY

πŸ”§ OUTILS INTERNES WEVAL

🧠 WEVIA Master
Chat IA souverain, 70+ intents, multi-agents
INTERNE
πŸ“Š L99 Brain
Dashboard L99, tests, NonReg, visual
INTERNE
πŸ—οΈ Architecture 3D
61 agents, 5 tiers, flux animΓ©s
INTERNE
⚑ AI Benchmark
Benchmark 14 providers, latence, coΓ»t
INTERNE
πŸ” DeerFlow Research
LangGraph deep research souverain
INTERNE
πŸ€– Agents Fleet
13 agents LIVE, monitoring
INTERNE
πŸ’¬ WEVIA Console
Console debug IA avancΓ©e
INTERNE
πŸ“ˆ Command Center
312 OK, 34 AUTH, 58 ERR monitoring
INTERNE
🗺 Cartographie live 3914 ecrans