Skip to content

IronMonitor 7.0.15

Choose a tag to compare

View all tags
@admercs admercs released this 02 Oct 21:53
· 230 commits to master since this release
v7.0.15
b84dd79

Add MLX-LM as an external Apple silicon inference provider for IronMonitor on macOS. The AI dropdown offers MLX (Apple Silicon), reads the server model list, and sends native monitoring tool calls through the validated read-only agent runtime. Automatic discovery supports CLI and TUI use; default_model is the MLX-LM startup-model alias. IronWorks remains preferred when both are available.

IronMonitor's MLX setup uses http://localhost:8082/v1 to avoid the IronWorks default port. On your Mac, install mlx-lm in a Python environment and start a compatible model:

mlx_lm.server --model <model-path-or-repo> --host 127.0.0.1 --port 8082

Close IronMonitor before upgrading:

cargo install ironmonitor --version 7.0.15 --locked --force
imon gui

Select MLX (Apple Silicon) on the AI page and choose the served model or default_model. See AI_INTEGRATION.md for installation and alternate-port library configuration. Native tool calling requires support from the server and chosen model/tokenizer.

Regression coverage exercises MLX model listing, native tool calls, returned tool evidence, default-model configuration, and GUI selection. Live MLX GPU inference is not tested on the Windows development host. Source release with the verified crate archive and SHA-256 checksum attached.

Assets 4
Loading