afm (llm/afm) Updated: 1 week ago Add to my watchlist
Local OpenAI-compatible Apple Foundation Models and MLX serverAFM provides a local AI server and command-line interface with Apple Foundation Models, MLX models, and a bundled WebUI.
Version: 0.9.19 License: MIT
GitHub
| Maintainers | i0ntempest |
| Categories | llm |
| Homepage | https://github.com/scouzi1966/maclocal-api |
| Platforms | {darwin >= 25} |
| Variants | - |
"afm" depends on
build (3)
Ports that depend on "afm"
No ports
Port notes
The startup item reads ${prefix}/etc/afm/afm_env.conf and
${prefix}/etc/afm/afm_args.conf. Edit those files, then restart the service.
Each non-comment line in the args file is a command or flag, optionally
followed by one value on the same line. With no command, afm starts its
default service on localhost:9999.
The service must run as a user with Apple Intelligence enabled. The wrapper script
uses AFM_USER to enter that user's launchd context and runs the service as
that user. Set a username, or a numeric UID prefixed with #. The default
value #501 is usually (but not guaranteed) the primary user's UID.
Change it as needed. A startup item has been generated that will aid in starting afm with launchd. It is disabled by default. Execute the following command to start it, and to cause it to launch at startup:
sudo port load afm
Port Health:
Loading Port Health