Glossary term

Runtime

The execution layer that actually loads and runs a local model.

A runtime is the engine or service responsible for loading a model and serving responses locally. Category: Runtime / formats Related topics: 0 Related resources: 3

Why this term matters

Users often confuse chat apps, model files, web UIs, and runtimes. Separating them makes model and hardware choices easier.

Examples

A local execution layer loads model files and exposes them to a chat or API surface.
Different execution layers can change setup friction, memory behavior, and hardware compatibility.
A polished interface can hide execution details from beginners.

Pages where this term matters

Software

Ollama

Command-line local model runtime that gives developers a clean path from downloading a model to exposing it through a local API.

Good fit: Developers and builders who want a local runtime that can plug into tools, scripts, and web UIs

5 min Updated Apr 25, 2026

Software

llama.cpp

Foundational local inference project that underpins many desktop apps, benchmarks, and local model experiments.

Good fit: Builders who want to understand the lower-level runtime layer beneath many local AI tools

5 min Updated Apr 26, 2026

Software

Open WebUI

Self-hosted web interface that usually makes sense after a user wants more flexibility than a desktop app.

Good fit: Users who want a flexible local web UI around Ollama or other model providers

5 min Updated Apr 25, 2026