Local Model
An AI model running on your own server (e.g., via Ollama), keeping data private and costs near zero.
Why it matters
Local models solve two problems: privacy (data never leaves your infrastructure) and cost (inference is free after setup).
In practice
Ollama running llama3.2 on our Hetzner server handles 80-95% of AI tasks for $0. Data stays on European servers.
Related terms
Ollama
A tool for running AI models locally. Free, private, fast.
Fallback
An alternative approach when the primary method fails (e.g., Ollama fails, Claude API takes over).
Quantization
Reducing a model's numerical precision to decrease size, cost, and inference time.
Inference
The process of an AI model generating a response or prediction from input data.