6,598 open-source and SaaS tools, with GitHub stats refreshed every day.

llamafile

Open source

llamafile packages a large language model and its runtime into one executable file that runs locally on most operating systems.

Open-source alternative to

docs.mozilla.ai
llamafile homepage screenshot
GitHub stars
26k
Last commit
yesterday
Repository age
3 years
Version
0.10.6
Licence
Custom
Self-hosted
Yes

About llamafile

llamafile is a Mozilla Builders project, now revamped by Mozilla.ai, that lets you distribute and run LLMs as one file. It combines llama.cpp with Cosmopolitan Libc into a framework that folds the complexity of running a model into one executable, known as a llamafile, which needs no installation and runs on your machine under most operating systems and CPU architectures. The goal is to make open models more approachable for developers and end users. The repository metadata lists the license as Other.

The project also bundles whisperfile, which applies the same single-file packaging to whisper.cpp for speech-to-text, transcribing and translating audio on the same platforms. Starting with version 0.10.0, llamafile uses a new build system that keeps it aligned with current llama.cpp, adding support for newer models while some earlier features may be missing. Previous versions remain on the releases page, and pre-built llamafiles state which server version they bundle.

Key features

  • Model and runtime in a single executable
  • Runs locally with no installation
  • Built on llama.cpp and Cosmopolitan Libc
  • Works across most OS and CPU architectures
  • whisperfile for speech-to-text and translation
  • Older versions kept on the releases page

Good fit for

  • →Sharing a local LLM as one downloadable file
  • →Offline transcription of audio files
Built with
C++
Tags
llm
local-ai
llama-cpp
single-file
gguf
speech-to-text
mozilla
offline

Open-source alternatives to llamafile

See all

SaaS alternatives to llamafile

See all