On-Prem · 36+ Agents IA · Souverainete des donnees On-Prem · 36+ AI Agents · Data Sovereignty

L'IA d'entreprise,
sur vos serveurs Enterprise AI,
on your servers

Deployez des agents IA pre-construits et personnalises directement sur votre infrastructure. LLMs locaux via Ollama, zero cout API, souverainete totale des donnees. Deploy pre-built and custom AI agents directly on your infrastructure. Local LLMs via Ollama, zero API costs, full data sovereignty.

Demander une demo Request a Demo Explorer les agents Explore Agents </> Guide d'installation Install Guide
$ git clone https://github.com/zeniclaw/core.git
$ cd zeniclaw && bash install.sh
# 5 containers: app, db, redis, gateway, ollama
$ open http://localhost:8080
36+
Agents IA
AI Agents
100%
On-Prem
13+
Modeles LLM
LLM Models
5 min
Deploiement
Deploy Time
// solutions

Trois piliers pour votre IA Three pillars for your AI

Une plateforme complete qui s'adapte a vos besoins : agents pre-construits, agents personnalises, et LLMs locaux. A complete platform that adapts to your needs: pre-built agents, custom agents, and local LLMs.

🤖

Agents pre-construits

Pre-built Agents

36+ agents IA specialises prets a l'emploi : revue de code, gestion de projet, finance, reunions, support RH, analyse de documents. Deployes en un clic, operationnels immediatement.

36+ specialized AI agents ready to use: code review, project management, finance, meetings, HR support, document analysis. Deployed in one click, operational immediately.

🧠

Agents personnalises (RAG)

Custom Agents (RAG)

Creez vos propres agents entraines sur vos documents internes. Retrieval-Augmented Generation pour des reponses precises basees sur vos donnees d'entreprise. Base de connaissances privee.

Create your own agents trained on your internal documents. Retrieval-Augmented Generation for precise answers based on your company data. Private knowledge base.

🖥

LLMs On-Prem (Ollama)

On-Prem LLMs (Ollama)

Executez Qwen, Mistral, Llama localement via Ollama. Zero cout API, aucune donnee ne quitte vos serveurs. Mode hybride disponible : Claude et GPT-4 quand necessaire.

Run Qwen, Mistral, Llama locally via Ollama. Zero API costs, no data leaves your servers. Hybrid mode available: Claude and GPT-4 when needed.

// use cases

L'IA dans chaque departement AI across every department

Des agents IA specialises pour chaque equipe de votre organisation. Specialized AI agents for every team in your organization.

👥 Ressources HumainesHuman Resources

  • Assistant d'onboarding pour les nouveaux employes
  • Onboarding assistant for new employees
  • Agent FAQ interne (politiques, avantages, conges)
  • Internal FAQ agent (policies, benefits, leave)
  • Suivi du bien-etre et de l'engagement
  • Wellness and engagement tracking

💰 Finance

  • Suivi des depenses et rapports automatises
  • Expense tracking and automated reports
  • Alertes budgetaires en temps reel
  • Real-time budget alerts
  • Analyse financiere et tableaux de bord
  • Financial analysis and dashboards

⚙ IT / DevOps

  • Revue de code automatisee et suggestions
  • Automated code review and suggestions
  • Reponse aux incidents et diagnostic
  • Incident response and diagnostics
  • Documentation technique automatique
  • Automatic technical documentation

⚖ JuridiqueLegal

  • Analyse de documents et contrats
  • Document and contract analysis
  • Verification de conformite reglementaire
  • Regulatory compliance checks
  • Recherche juridique assistee par IA
  • AI-assisted legal research

📈 CommercialSales

  • Assistant CRM et suivi des opportunites
  • CRM assistant and opportunity tracking
  • Qualification automatique des leads
  • Automated lead qualification
  • Preparation de propositions commerciales
  • Sales proposal preparation

📋 OperationsOperations

  • Comptes rendus de reunion automatiques
  • Automatic meeting notes and summaries
  • Gestion de projet et suivi des taches
  • Project management and task tracking
  • Automatisation des processus metier
  • Business process automation
// agents

36+ agents IA specialises 36+ specialized AI agents

Chaque agent dispose de son propre systeme de prompt, acces aux outils, et memoire isolee. Le RouterAgent dispatch intelligemment chaque requete. Each agent has its own system prompt, tool access, and isolated memory. The RouterAgent intelligently dispatches every request.

💬

ChatAgent

General conversation & multimodal

💻

DevAgent

Code, GitHub & API automation

⏰

ReminderAgent

Time-based reminders & snooze

📁

ProjectAgent

Project management & stats

📈

AnalysisAgent

Data analysis & report generation

✅

TodoAgent

Task management & priorities

🎵

MusicAgent

Spotify search & playlists

💰

FinanceAgent

Expenses, budgets & alerts

😊

MoodCheckAgent

Emotional tracking & wellness

🔍

CodeReviewAgent

Intelligent code analysis

🧠

SmartContextAgent

User behavior learning

🎙

SmartMeetingAgent

Meeting notes & transcription

🎤

VoiceCommandAgent

Audio transcription & commands

📚

FlashcardAgent

Spaced repetition learning

📅

EventReminderAgent

Calendar events & scheduling

📰

ContentSummarizerAgent

URL & content summarization

📷

ScreenshotAgent

OCR, annotation & comparison

🎯

HabitAgent

Habit tracking & streaks

🍅

PomodoroAgent

Focus timer & productivity

📄

DocumentAgent

PDF, Excel & document generation

🎮

HangmanGameAgent

Word games & entertainment

🔍

WebSearchAgent

Real-time web search & API stats

💳

BudgetTrackerAgent

Personal finance & budgets

📖

ContentCuratorAgent

Reading lists & challenges

🧩

InteractiveQuizAgent

Dynamic quizzes & scoring

📧

DailyBriefAgent

Daily briefing compilation

📌

ContextAgent

Context memory management

💭

ConversationMemoryAgent

Conversation history tracking

🍳

RecipeAgent

Recipe recommendations

🎲

GameMasterAgent

RPG game master

🤝

CollaborativeTaskAgent

Team task coordination

⏳

TimeBlockerAgent

Time blocking & scheduling

⚙

UserPreferencesAgent

Preference management

📋

ZenibizDocsAgent

Photo to PDF conversion

🤖

AIAssistantAgent

Self-improving AI assistant

🔀

RouterAgent

Confidence-scored auto-dispatch

// custom agents

Creez vos propres agents metier Build your own business agents

Entrainez des agents IA sur vos documents internes grace au RAG. Vos employes obtiennent des reponses precises basees sur votre base de connaissances. Train AI agents on your internal documents using RAG. Your employees get precise answers based on your knowledge base.

📂

Importez vos documents

Import your documents

PDF, Word, Excel, pages web. Indexation automatique et vectorisation pour une recherche semantique rapide.

PDF, Word, Excel, web pages. Automatic indexing and vectorization for fast semantic search.

🎯

Reponses contextuelles

Contextual answers

Retrieval-Augmented Generation : l'agent cite ses sources et ne fabrique jamais de reponses. Regles anti-hallucination integrees.

Retrieval-Augmented Generation: the agent cites its sources and never fabricates answers. Built-in anti-hallucination rules.

🔒

100% prive

100% private

Vos documents restent sur vos serveurs. Combinez avec les LLMs on-prem pour une souverainete totale, sans aucun appel API externe.

Your documents stay on your servers. Combine with on-prem LLMs for full sovereignty, with zero external API calls.

// technology

Architecture enterprise-grade Enterprise-grade architecture

Choisissez votre mode de deploiement. Full on-prem, full cloud ou hybride — la meme plateforme, votre infrastructure. Choose your deployment mode. Full on-prem, full cloud or hybrid — same platform, your infrastructure.

graph LR subgraph Channels["Canaux"] WA["WhatsApp
Baileys"] WEB["Web Chat
Dashboard"] API["API REST"] end subgraph STACK["ZeniClaw Stack — Vos Serveurs"] GW["WhatsApp
Gateway
:3000"] APP["App Laravel 12
PHP 8.4
:8080"] ROUTER["RouterAgent
Intent Classifier"] AGENTS["36+ Agents
Agentic Loop
RAG Engine"] CUSTOM["Custom Agents
Skills & Scripts"] DB[("PostgreSQL 16
Vectors + Memory")] CACHE[("Redis 7
Queue + Cache")] end subgraph LLM["LLMs On-Prem"] OLLAMA["Ollama
Qwen 2.5 | Mistral
Llama | DeepSeek"] end WA <-->|QR| GW WEB --> APP API --> APP GW -->|webhook| APP APP --> ROUTER ROUTER --> AGENTS AGENTS --> CUSTOM AGENTS <-->|embeddings| OLLAMA CUSTOM <-->|inference| OLLAMA APP --- DB APP --- CACHE style STACK fill:#0a2e1a,stroke:#10b981,stroke-width:2px style LLM fill:#0a2e1a,stroke:#10b981,stroke-width:2px style Channels fill:#1a1f2e,stroke:#334155 style OLLAMA fill:#065f46,stroke:#10b981,color:#fff style APP fill:#1e3a5f,stroke:#3b82f6,color:#fff style DB fill:#1e3a5f,stroke:#3b82f6,color:#fff style CACHE fill:#0f3a2a,stroke:#10b981,color:#fff
graph LR subgraph Channels["Canaux"] WA["WhatsApp
Baileys"] WEB["Web Chat
Dashboard"] API["API REST"] end subgraph STACK["ZeniClaw Stack — Vos Serveurs"] GW["WhatsApp
Gateway
:3000"] APP["App Laravel 12
PHP 8.4
:8080"] ROUTER["RouterAgent
Intent Classifier"] AGENTS["36+ Agents
Agentic Loop
RAG Engine"] CUSTOM["Custom Agents
Skills & Scripts"] DB[("PostgreSQL 16
Vectors + Memory")] CACHE[("Redis 7
Queue + Cache")] end subgraph LOCAL["LLMs On-Prem"] OLLAMA["Ollama
Qwen 2.5 | Mistral"] end subgraph CLOUD["Cloud LLMs"] CLAUDE["Claude
Opus | Sonnet | Haiku"] GPT["OpenAI
GPT-4o"] end WA <-->|QR| GW WEB --> APP API --> APP GW -->|webhook| APP APP --> ROUTER ROUTER --> AGENTS AGENTS --> CUSTOM AGENTS <-->|"fast tasks"| OLLAMA AGENTS <-->|"complex reasoning"| CLAUDE AGENTS -.->|optional| GPT APP --- DB APP --- CACHE style STACK fill:#111827,stroke:#3b82f6,stroke-width:2px style LOCAL fill:#0a2e1a,stroke:#10b981,stroke-width:2px style CLOUD fill:#1a1035,stroke:#8b5cf6,stroke-width:2px style Channels fill:#1a1f2e,stroke:#334155 style OLLAMA fill:#065f46,stroke:#10b981,color:#fff style CLAUDE fill:#3b1f7e,stroke:#8b5cf6,color:#fff style GPT fill:#3b1f7e,stroke:#8b5cf6,color:#fff style APP fill:#1e3a5f,stroke:#3b82f6,color:#fff style DB fill:#1e3a5f,stroke:#3b82f6,color:#fff style CACHE fill:#0f3a2a,stroke:#10b981,color:#fff
graph LR subgraph Channels["Canaux"] WA["WhatsApp
Baileys"] WEB["Web Chat
Dashboard"] API["API REST"] end subgraph STACK["ZeniClaw Stack — Vos Serveurs"] GW["WhatsApp
Gateway
:3000"] APP["App Laravel 12
PHP 8.4
:8080"] ROUTER["RouterAgent
Intent Classifier"] AGENTS["36+ Agents
Agentic Loop
RAG Engine"] CUSTOM["Custom Agents
Skills & Scripts"] DB[("PostgreSQL 16
Vectors + Memory")] CACHE[("Redis 7
Queue + Cache")] end subgraph CLOUD["Cloud AI Providers"] CLAUDE["Claude
Opus | Sonnet | Haiku"] GPT["OpenAI
GPT-4o | GPT-4o-mini"] EMBED["Embeddings
text-embedding-3"] end WA <-->|QR| GW WEB --> APP API --> APP GW -->|webhook| APP APP --> ROUTER ROUTER --> AGENTS AGENTS --> CUSTOM AGENTS <-->|inference| CLAUDE AGENTS <-->|fallback| GPT AGENTS <-->|RAG| EMBED APP --- DB APP --- CACHE style STACK fill:#111827,stroke:#3b82f6,stroke-width:2px style CLOUD fill:#1a1035,stroke:#8b5cf6,stroke-width:2px style Channels fill:#1a1f2e,stroke:#334155 style CLAUDE fill:#3b1f7e,stroke:#8b5cf6,color:#fff style GPT fill:#3b1f7e,stroke:#8b5cf6,color:#fff style EMBED fill:#3b1f7e,stroke:#8b5cf6,color:#fff style APP fill:#1e3a5f,stroke:#3b82f6,color:#fff style DB fill:#1e3a5f,stroke:#3b82f6,color:#fff style CACHE fill:#1e3a5f,stroke:#3b82f6,color:#fff
🐘
PHP 8.4
🏰
Laravel 12
⚡
Livewire 3
🎨
Tailwind CSS
🗃
PostgreSQL 16
🔴
Redis 7
🐳
Podman / Docker
📱
Baileys Gateway
🧠
Claude / GPT-4
🖥
Ollama
// deploy

Operationnel en 5 minutes Up and running in 5 minutes

Un serveur Linux avec Podman ou Docker. Le script d'installation gere tout le reste. A Linux server with Podman or Docker. The install script handles everything else.

1

Clonez le depot

Clone the repository

Recuperez le code source et lancez l'installateur interactif.

Fetch the source code and run the interactive installer.

git clone https://github.com/zeniclaw/core.git ~/zeniclaw
cd ~/zeniclaw
2

Lancez l'installation

Run the installer

Le script configure automatiquement les 5 conteneurs (app, base de donnees, cache, gateway, LLM).

The script automatically configures all 5 containers (app, database, cache, gateway, LLM).

sudo bash install.sh
# 5 containers: app, postgres, redis, gateway, ollama
3

Ouvrez votre navigateur

Open your browser

Accedez au tableau de bord, configurez vos agents et connectez vos canaux (WhatsApp, Web Chat, API).

Access the dashboard, configure your agents and connect your channels (WhatsApp, Web Chat, API).

open http://localhost:8080
# Default: admin@zeniclaw.io / password
// security

Securite et conformite Security & compliance

Concu pour les exigences de securite des entreprises. Vos donnees ne quittent jamais vos serveurs. Built for enterprise security requirements. Your data never leaves your servers.

🔒

Souverainete des donnees

Data sovereignty

Deploiement 100% on-prem. Aucune donnee ne transite par des serveurs tiers. Compatible air-gap avec LLMs locaux.

100% on-prem deployment. No data passes through third-party servers. Air-gap compatible with local LLMs.

🔐

Chiffrement AES-256

AES-256 Encryption

Tous les secrets et donnees sensibles sont chiffres au repos. Communications HTTPS/TLS en transit.

All secrets and sensitive data encrypted at rest. HTTPS/TLS communications in transit.

👥

RBAC

Controle d'acces base sur les roles : superadmin, admin, operateur, lecteur. Permissions granulaires par agent.

Role-based access control: superadmin, admin, operator, viewer. Granular permissions per agent.

📝

Journaux d'audit

Audit logs

Tracabilite complete de chaque action, decision d'agent, et acces utilisateur. Export pour conformite.

Complete traceability of every action, agent decision, and user access. Export for compliance.

🛡

Sandboxing des agents

Agent sandboxing

Chaque agent s'execute dans un environnement isole avec memoire et outils dedies. Aucun acces croise non autorise.

Each agent runs in an isolated environment with dedicated memory and tools. No unauthorized cross-access.

🔄

Mises a jour et monitoring

Updates & monitoring

Mises a jour automatiques, health checks, et watchdog integre pour une disponibilite maximale.

Automatic updates, health checks, and built-in watchdog for maximum uptime.

// contact

Parlons de votre projet Let's discuss your project

Une question, une demo, un projet d'integration ? Contactez-nous. A question, a demo, an integration project? Get in touch.

Guillaume Tilleul

Guillaume Tilleul

Fondateur — ZeniBiz Founder — ZeniBiz

Entrepreneur et architecte de solutions IA pour les entreprises. Passionné par le déploiement on-prem et la souveraineté des données. Entrepreneur and AI solution architect for businesses. Passionate about on-prem deployment and data sovereignty.

✉ gti@zenibiz.com 📱 +32 484 88 58 71 🌐 zenibiz.com
💬
Communaute WhatsApp ZeniClaw ZeniClaw WhatsApp Community
Rejoignez la communaute — aide, idees, mises a jour Join the community — help, ideas, updates
→