01 / Infrastructure · Local inference
Home Lab & Self-Hosted AI
A private cloud across three servers running five containerized workloads with automated updates, hourly health checks, and self-healing recovery. Two local LLMs on Ollama and llama.cpp drive six MCP tools that automate my daily workflows, all reachable securely over TLS and a VPN.
- 3
- Servers
- 5
- Workloads
- 64
- Tokens / sec
- Linux
- Docker
- Ollama
- llama.cpp
- MCP
- Prometheus
- Grafana