Local & Uncensored · verified 2026-07-27

llama.cpp

The inference engine everything else is built on.

Georgi Gerganov's C/C++ implementation of LLM inference. Quantised GGUF models run on laptops, phones and servers, and most local tools embed it under the hood.

OPEN LLAMA.CPP

Rating: 4.7 / 5
Pricing: Open source
Best for: Self-hosting, Offline, Cost control
Platforms: Linux, macOS, Windows, CLI, API
Runs locally: Yes
Content policy: Standard filters

Matching filters

Research Model hosting & inference Has a public API Works offline Open source Free or open source Free when self-hosted macOS Windows Linux Desktop app CLI / terminal API Runs locally

What stands out

GGUF quantisation
CPU, CUDA, Metal and Vulkan back ends
MIT licence

Also in Local & Uncensored

ComfyUI

Node graphs for people who want control.

Ollama

One command to run a model locally.

LM Studio

A desktop app for local models.