Disclosure: this article contains affiliate links. If you buy through them we may earn a commission at no additional cost to you. Our recommendations are based on the evidence recorded alongside this article, and prices were verified on the date shown.

What changed and why it matters

A project called What the Model — wtm for short — is listed on GitHub as a Local LLM Inventory for macOS, published at github.com/powtac/wtm. That's the entire premise stated on the project's own page, and it's specific enough to verify yourself in one click: open the repository and the description reads exactly that.

The name signals a narrow job. It's not billed as a model runner, a chat client, or a fine-tuning toolkit — it's billed as inventory, which points to tracking and organizing local model files rather than executing them. If you've pulled weights from more than one source, you already know they don't sort themselves into one place on disk, and a tool built purely to solve that is a sign of how granular local AI tooling on macOS has gotten.

The detail

According to the wtm GitHub listing, here is what's stated:

Everything in that table comes directly from the repository's own listing, and nothing in it is inferred. wtm doesn't describe itself as software that loads or serves models — the word doing the work is "inventory," which is a tracking function, not a runtime.

What the listing doesn't include, at least in the material available here, is a published feature list, version history, system requirements, or pricing. That's worth saying plainly rather than filling in with a guess: if you're deciding whether wtm fits your setup, the GitHub repository is the primary source to check before installing anything, since it's the only place those specifics would be documented and kept current.

Skip it if: you're not on macOS, since the project is positioned specifically for that platform, or you're looking for something that runs or serves models rather than tracking which ones are already on your disk.

What it means for you

If you're managing local LLM files on a Mac, a tool positioned like wtm answers one question: what do you actually have installed. It doesn't answer the next question most people run into once inventory is sorted — where you actually run inference, put an app in front of a model, or scale past what one machine can handle.

That second question is where AI infrastructure platforms come in, and it's a different category from a local inventory tool. Vercel operates in that AI tools space, covering deployment and hosting rather than tracking files on your own drive. [See current Vercel pricing](See current pricing at the merchant) to compare plan tiers before you decide whether to keep your model workflow entirely local, move parts of it to a hosted platform, or run both side by side.

Before committing budget to either path, it helps to have a consistent way to compare cost structures instead of eyeballing them. Our guide on MI Measure Price Calculator: How to Choose walks through how to line up different pricing models against each other, so a flat local setup and a metered cloud bill get compared on the same terms instead of on gut feel.

Best for: developers who already keep multiple local model files on macOS and want a dedicated way to track them, and who are also weighing whether a hosted platform makes sense for the deployment half of the workflow.

Skip it if: your models already live entirely in the cloud, or you're not running anything locally that needs inventorying in the first place — in that case, the deployment platform is the whole decision, not a follow-on to it.