Alternatives to Foundry Local
Microsoft's tool for downloading and running AI models entirely on your own device. The listings below can replace it for an important use case. Each note says what changes if you switch.
The original
Foundry Local
Microsoft's tool for downloading and running AI models entirely on your own device.
Replacements
Listings that take over the same core job as Foundry Local.
Ollama
The simplest way to pull down an open language model and run it on your own machine, from one command or a desktop app.
Ollama is MIT licensed, runs on Windows, macOS and Linux, and exposes a localhost REST API, though the company also sells cloud inference.
LM Studio
A polished desktop app for downloading, running and chatting with local language models, with an OpenAI-compatible server built in.
LM Studio adds a desktop interface for downloading models from Hugging Face and serves OpenAI-compatible endpoints, with Python and TypeScript SDKs, and is also closed source.
Lemonade
A local AI runtime that serves text, image and speech models through a GUI, CLI and API.
Lemonade serves text, image and speech models through a GUI, CLI and API, states zero telemetry, and adds Linux packages.
llama.cpp
The C++ engine most local AI apps are built on, running language models on ordinary hardware.
llama.cpp is an MIT-licensed C++ engine with a built-in server and many GPU backends, running on Linux and Android as well as Windows and macOS.
LocalAI
Serve language, image and speech models on your own hardware.
LocalAI is an MIT-licensed self-hosted server with compatible APIs for common AI services, running on Linux and macOS rather than Windows.
llamafile
Packages a language model and the code to run it into one executable file that works on several operating systems without installation.
llamafile packages a model and runtime into one Apache-licensed executable that runs across Windows, macOS, Linux and BSD without installation.
RamaLama
Command-line tool that pulls AI models from any source and serves them locally in containers.
RamaLama is an MIT-licensed command-line tool that serves models in containers on macOS and Linux, requiring a container engine such as Podman or Docker.
Xinference
An open-source inference server for running language, speech and multimodal models through one API.
Xinference is an open-source server offering a GPT-style drop-in API for language, speech and multimodal models, on macOS and Linux and scaling to servers.
Similar software
Related functionality, not a direct replacement.
GAIA
AMD's open-source app for building and running local AI agents on Ryzen AI PCs.
AI Playground
Intel's desktop app for local AI image creation, image stylizing and chatbot use on Arc GPUs.
vLLM
High-throughput, memory-efficient inference and serving engine for LLMs.
Open WebUI
A self-hosted chat interface that talks to Ollama and any OpenAI-compatible API, so your local models get a proper front end.
Jan
An open-source ChatGPT replacement that runs entirely offline, and can also front cloud providers when you want them.
LLM
A command-line tool and Python library for prompting remote and locally installed language models.
