⚡ Hosted on Cloudflare Pages Global Edge

Intelligent Edge Systems & Private Model Infrastructure

A high-performance static portal deployed at the Cloudflare edge, interconnecting local Strix Halo unified-inference clusters, autonomous agents, and persistent research knowledge graphs.

Explore Services API Docs & cURL

131,072

Inference Context

< 15 ms

Cloudflare Edge TTFB

100%

Vault Date Verified

128 GB

Unified APU Memory

Core Infrastructure

Cluster Services & Capabilities

Strata Inference Engine

Port 8080 · 131k

High-throughput local LLM engine powering qwen3.8-flash-next with full 131k token context on AMD Strix Halo unified LPDDR5X.

48 tok/s · Vulkan/ROCm CT 100

Hermes & Obsidian Vault

301/301 Entities

Automated daily research synthesis and knowledge graph. Fully audited with set-once provenance dates and zero dead wikilinks.

Git bare sync · 30m cron CT 101

SearXNG Metasearch

Port 8888

Decentralized, privacy-preserving metasearch aggregator querying across Brave, Google, Bing, and Wikipedia with zero tracking.

JSON API · High QPS CT 102

Honcho Agent Memory

Port 8000

Cross-session representation, dynamic conclusions, and persistent multi-agent user cards for personalized intelligence.

FastAPI · Deriver daemon CT 101

ms-06s Proxmox Host

SP5100 HW Watchdog

AMD Ryzen AI MAX+ 395 workstation with physical MMIO hardware watchdog, EarlyOOM protection, and 23.5 GiB buffer guarantees.

123 GiB RAM · 32 Threads Host 192.168.8.99

Fileshare & Git Server

SMB / Z: Share

Centralized NVMe ZFS storage pool, multi-agent workspaces, and bare Git repositories serving all workstation nodes.

ZFS pool · 10GbE network CT 104

Integration

API Endpoints & cURL Testing

Topology

Network Architecture Map

Node / Container	IP Address	Primary Role	Resources	Status
ms-06s (Host)	`192.168.8.99`	Proxmox VE 8.4 Host / AMD Strix Halo	16C/32T · 128 GB LPDDR5X	Online
CT 100 (llm)	`192.168.8.100`	Strata LLM (131k context) + Ollama	100 GiB CGroup · Radeon 8060S	Active
CT 101 (hermes)	`192.168.8.101`	Hermes Agent, Knowledge Vault, Honcho	8 Cores · 16 GB RAM	Active
CT 102 (searxng)	`192.168.8.102`	SearXNG Privacy Metasearch Aggregator	4 Cores · 8 GB RAM	Active
CT 104 (fileshare)	`192.168.8.104`	Central SMB Share (Z:) & Bare Git Repos	8 Cores · 32 GB RAM · NVMe ZFS	Active
CT 105 (telegram)	`192.168.8.105`	Telegram Gateway & Live Assistant	4 Cores · 8 GB RAM	Active