Guides
Self-hosted AI guides
Practical setup guides and comparisons for running Ollama, LiteLLM, and open-source AI models on your own infrastructure.
Cheapest VPS for Ollama in 2026
We tested 5 providers. Hetzner CX22 at €5.89/mo wins on pure value. Here's the full comparison with real community benchmarks.
I Replaced GitHub Copilot with a $8/mo VPS. Here's How.
Qwen 2.5 Coder 7B on Hetzner + Continue.dev in VS Code. Private code completion with no rate limits. Full setup walkthrough.
LiteLLM + Ollama: Route 80% of Traffic Away From the API
How to run a LiteLLM proxy that sends cheap tasks to your local Ollama instance and only falls back to Claude/GPT when needed. Monthly bill: $22 instead of $180.
Which Models Can Your Hardware Actually Run? A RAM Guide
The practical RAM requirements for every popular model size — 3B through 70B. Which VPS tier or hardware do you actually need?
Mac Mini M4 for Ollama: 3 Months of Real Use
We ran a Mac Mini M4 16GB as an always-on AI server for 3 months. Real power costs, real benchmarks, and an honest verdict on when it's worth it.