llamafile
Run LLMs with a single executable file
24.2k|stable|0.10.0
Linux
Single-file distribution — Package LLMs as standalone executables with no installation required.
Cross-platform support — Run on Windows, macOS, Linux, and BSD with the same binary.
CPU optimization — Built on llama.cpp for efficient inference across hardware architectures.
Speech-to-text — Includes whisperfile for transcription and translation without dependencies.
llamafile packages open LLMs into single-file executables that run locally on any OS with zero setup. Built on Mozilla's Cosmopolitan Libc, it eliminates the complexity of model installation and GPU requirements. Includes whisperfile for offline speech-to-text on the same portable model.
Source on GitHub
pavilion