llamafile
A whole chatbot in one portable file
What it is
Packages an LLM's weights and the llama.cpp engine into a single cross-platform executable using Cosmopolitan Libc. One downloaded file runs the same model on macOS, Linux, Windows, and the BSDs — no install, no runtime, no dependencies. A sibling, whisperfile, does the same for speech-to-text.
Why it's interesting
The most radical answer to AI distribution anyone has shipped: a model as a file you can archive, air-gap, or hand someone on a USB stick. Revived with fresh releases under Mozilla.ai after a quiet spell.
Use cases
- Offline and air-gapped LLM chat
- Distributing a model plus runtime as one artifact
- An embedded OpenAI-compatible local server
Who it's for
Developers, sysadmins, privacy-focused tinkerers, digital preservationists
Setup
Trivial. Download one file, chmod +x, run. Any modern CPU; GPU optional
Limitations & cautions
Windows caps executables at 4GB, so larger models need external weights. Feature-wise it trails upstream llama.cpp and has fewer knobs than dedicated servers.
Editorial takeaway
Fifty years from now, when someone wants to run 2020s AI, this is the format that will still boot.