# llamafile — Single-File LLM, No Install Needed

> llamafile distributes LLMs as single-file executables that run on any OS. 23.9K+ GitHub stars. No installation, cross-platform, built on llama.cpp + Cosmopolitan. Apache 2.0.

## Install

Save in your project root:

## Quick Use

```bash
# Download and run (no install!)
curl -LO https://huggingface.co/mozilla-ai/llamafile_0.10.0/resolve/main/Qwen3.5-0.8B-Q8_0.llamafile
chmod +x Qwen3.5-0.8B-Q8_0.llamafile
./Qwen3.5-0.8B-Q8_0.llamafile

# Opens a web UI at http://localhost:8080
# Also includes whisperfile for speech-to-text
```

Works on macOS, Linux, Windows, FreeBSD — same file, no dependencies.

---

## Intro

llamafile enables distributing and running large language models as single-file executables that work across multiple operating systems and CPU architectures with zero installation. With 23,900+ GitHub stars and Apache 2.0 license, it combines llama.cpp with Cosmopolitan Libc to create portable, installationless applications. Download one file, make it executable, run it — instant AI on any platform. Also includes whisperfile for speech-to-text in the same single-file format.

**Best for**: Anyone who wants the absolute simplest way to run an LLM — one file, no setup
**Works with**: Claude Code, OpenAI Codex, Cursor, Gemini CLI, Windsurf
**Platforms**: macOS, Linux, Windows, FreeBSD (same binary)

---

## Key Features

- **Single file**: Entire LLM + runtime in one executable
- **Zero install**: No Python, no Docker, no dependencies
- **Cross-platform**: Same file runs on macOS, Linux, Windows, FreeBSD
- **Built-in web UI**: Opens localhost:8080 with chat interface
- **whisperfile**: Speech-to-text in the same single-file format
- **Built on llama.cpp**: Full model compatibility and performance
- **Cosmopolitan Libc**: Universal binary technology for portability

---

### FAQ

**Q: What is llamafile?**
A: llamafile packages LLMs as single-file executables with 23.9K+ stars. No installation — download, chmod +x, run. Works on macOS/Linux/Windows/FreeBSD. Apache 2.0.

**Q: How do I use llamafile?**
A: Download a .llamafile from HuggingFace, `chmod +x` it, and run it. A web UI opens at localhost:8080. No other setup needed.

---

## Source & Thanks

> Created by [Mozilla](https://github.com/Mozilla-Ocho). Licensed under Apache 2.0.
> [Mozilla-Ocho/llamafile](https://github.com/Mozilla-Ocho/llamafile) — 23,900+ GitHub stars

---
Source: https://tokrepo.com/en/workflows/f30a667e-1b6a-4140-a92f-b5027f16c8bc
Author: AI Open Source