gandalf/secubox-openwrt
1
1
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
e7a9062140
secubox-openwrt/package/secubox/secubox-app-localai/README.fr.md
CyberMind-FR ccfb58124c docs: Add trilingual documentation (French and Chinese translations) 
Add complete French (fr) and Chinese (zh) translations for all documentation:

- Root files: README, CHANGELOG, SECURITY, BETA-RELEASE
- docs/: All 16 core documentation files
- DOCS/: All 19 deep-dive documents including embedded/ and archive/
- package/secubox/: All 123+ package READMEs
- Misc: secubox-tools/, scripts/, EXAMPLES/, config-backups/, streamlit-apps/

Total: 346 translation files created

Each file includes language switcher links for easy navigation between
English, French, and Chinese versions.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2026-03-20 10:00:18 +01:00

3.2 KiB
Raw Blame History

English | Francais | 中文

SecuBox LocalAI

Serveur LLM natif avec API REST compatible OpenAI. Supporte les modeles GGUF sur ARM64 et x86_64.

Version : 3.9.0

Fonctionnalites

  • API REST compatible OpenAI (/v1/chat/completions, /v1/completions, /v1/embeddings)
  • Support des modeles GGUF (LLaMA, Mistral, Phi, TinyLlama, etc.)
  • Architectures ARM64 et x86_64
  • Panneau Agent Jobs (v3.9) -- Planifiez et gerez les taches agentiques en arriere-plan
  • Reclamateur Memoire (v3.9) -- Eviction LRU pour les modeles charges
  • Support des embeddings (preset GTE-Small)

Installation

opkg install secubox-app-localai

# Telecharger le binaire (premiere execution)
localaictl install

# Installer un modele
localaictl model-install tinyllama

# Activer et demarrer le service
uci set localai.main.enabled=1
uci commit localai
/etc/init.d/localai enable
/etc/init.d/localai start

Configuration

Fichier de configuration UCI : /etc/config/localai

config localai 'main'
    option enabled '0'
    option api_port '8081'
    option api_host '0.0.0.0'
    option models_path '/srv/localai/models'
    option threads '4'
    option context_size '2048'

Presets de Modeles

Preset Taille Description
tinyllama 669Mo TinyLlama 1.1B (chat)
phi2 1.6Go Microsoft Phi-2 (chat)
mistral 4.1Go Mistral 7B Instruct (chat)
gte-small 67Mo GTE Small (embeddings) 
localaictl model-install tinyllama
localaictl model-install gte-small   # Pour les embeddings

Commandes CLI

# Installation/Desinstallation
localaictl install          # Telecharger le binaire depuis GitHub
localaictl uninstall        # Supprimer le binaire

# Service
localaictl start            # Demarrer le service
localaictl stop             # Arreter le service
localaictl restart          # Redemarrer le service
localaictl status           # Afficher le statut
localaictl logs [-f]        # Afficher les logs

# Modeles
localaictl models           # Lister les modeles installes
localaictl model-install <name>  # Installer un modele
localaictl model-remove <name>   # Supprimer un modele

# Backends
localaictl backends         # Lister les backends disponibles

Points d'Acces API

Port par defaut : 8081

Point d'acces Description
/v1/models Lister les modeles charges
/v1/chat/completions Completion de chat
/v1/completions Completion de texte
/v1/embeddings Generer des embeddings
/readyz Verification de sante

Fichiers

  • /etc/config/localai -- Configuration UCI
  • /usr/sbin/localaictl -- CLI du controleur
  • /usr/bin/local-ai -- Binaire (telecharge)
  • /srv/localai/models/ -- Stockage des modeles

Dependances

  • libstdcpp
  • libpthread
  • wget-ssl
  • ca-certificates

Integration SecuBox

LocalAI sert de backend d'inference pour SecuBox AI Gateway (Couche 2) :

  • Inference LLM locale en priorite
  • Integration MCP Server pour les outils agents
  • Support des agents Threat Analyst, CVE Triage et autres agents autonomes

Licence

MIT

Sources