Run llama.cpp from ChatGPT or Claude

• MacWorks

  Homebrew has a ready-made build; the AI can install it with brew install llama.cpp.

• LinuxWorks

  Homebrew on Linux has a ready-made build.

• WindowsNot in Homebrew

  Homebrew does not run on Windows. If llama.cpp has a Windows build, BCD can run it in PowerShell.

Version

0.9.0

License

MIT

Installs (last 365 days)

289,838 · #66 in Homebrew

Homepage

llama.app

Install

brew install llama.cpp

“Install llama.cpp with Homebrew and show me its version”
“Read llama.cpp --help and give me three useful examples for my files”
“Update llama.cpp and tell me what changed”

ggml · openssl@4

whisper.cpp · ramalama

Can ChatGPT run llama.cpp on my computer?
Yes. With BCD on your Mac or Linux computer and command permission turned on for it, ChatGPT runs llama.cpp there and reads the output. Claude and other apps with MCP connectors can too.

Do I need to install llama.cpp first?
No. If Homebrew is installed, the AI can run brew install llama.cpp for you.

Is it safe?
A new computer starts read-only: running commands is turned on per computer, and BCD reaches only the folders you allow.

Try it on your computer

BCD connects ChatGPT, Claude and other AI apps to your own Mac, Windows or Linux computer. One computer is free, with 20,000 tool calls a month and no card.