Llamafile packages open large language models into a single-file executable that runs locally across most operating systems and CPU architectures without installation. It combines llama.cpp with Cosmopolitan Libc to simplify distribution and execution of models and speech-to-text tools.
Project overview
It eliminates installation and cross-platform compatibility barriers by collapsing LLM complexity into a single executable that runs locally across most operating systems and CPU architectures.
Project type
Model Runtime
Use cases
Chat Assistants · Meeting Notes · Audio & Speech
Deployment
Refer to project documentation
License
License pending
Best for
Developers and general users seeking to run text generation or speech-to-text tasks locally from the command line without performing software installations.
Key capabilities
Collapses the complexity of LLMs into a single executable file that runs locally on most operating systems and CPU architectures without installation.
Provides a single-file speech-to-text tool built on whisper.cpp for audio transcription and translation across supported platforms without installation.
Limitations and risks
Only executables under 4GB can run on Windows directly without using external weights.
The v0.10.* series uses a new build system that might be missing some features from previous versions.
Getting started
Download the example model executable, make it executable (such as by using chmod +x), and run it from the command line.
Evidence and sources
GitHub project description: Distribute and run LLMs with a single file.
README: combining [llama.cpp](https://github.com/ggerganov/llama.cpp) with [Cosmopolitan Libc](https://github.com/jart/cosmopolitan) into one framework that collapses all the complexity o…
README: llamafile lets you distribute and run LLMs with a single file.
README: llamafile also includes **[whisperfile](https://docs.mozilla.ai/llamafile/whisperfile)**, a single-file speech-to-text tool built on [whisper.cpp](https://github.com/ggerganog/whi…
README: Download and run your first llamafile in minutes: