5b3d4e0e1d
- tools/organ.py: Pack/install/publish AI persona organs - tools/forge.sh: Convert HuggingFace models to GGUF, quantize - tools/store.sh: Browse/install/publish community models - tools/compilation/: Cross-platform build scripts (Linux/macOS) - scripts/install.sh: Universal installer (auto-detect OS/arch/GPU) - site/saas/: SaaS frontend + backend source (v3 unified design) - site/vitrine/: Main site source (inference-x.com) - docs/ARCHITECTURE.md: Full system architecture All plans now in test mode - Studio & Enterprise free to test. Branch: master | Maintainer: Anti-Atlas craton (elmadani)
1001 B
1001 B
IX System Architecture
Infrastructure
inference-x.com (ARCHE · OVH) build.inference-x.com (OASIS · Hetzner)
├── nginx reverse proxy ├── ix-saas (Node.js, PM2, port 4080)
├── Gitea (port 3000) ├── echo brain (port 8089)
└── Site vitrine (HTML) ├── invoke gateway (port 3001)
└── OneCloud VM pool (demos)
Demo Flow
Visitor → POST /api/demo/start → OneCloud VM (2vCPU/2GB) → IX engine → LLaMA 3.2 1B → SSE stream → Chat → 30min → VM destroyed
Plans (Test Mode)
- Community: Free forever, full local engine
- Studio Test: Free, 2vCPU cloud instance, store access
- Enterprise Test: Free, 8vCPU/32GB, forge, store publish
API
Public: /api/health, /api/demo/stats, /api/demo/start, /api/demo/stream/:token, /api/community/scout Auth: /api/auth/register, /api/auth/login, /api/builds, /api/instance/provision, /api/store