Inferix is a model hub and a GPU marketplace in one place. You can host models, publish datasets, ship Spaces, deploy inference endpoints, and rent GPU time by the second — without moving your work between four different providers. This blog is where we write down how the platform actually behaves: what a feature does, what it costs, and where the edges are. It is written by the team building it. ## What is here today The hub holds roughly 11,900 model repositories and 4,400 datasets, browsable and searchable without an account. Model pages carry cards, file trees, and hardware-compatibility information for quantized builds. Datasets get a browser-based viewer with column statistics and a SQL console. On the compute side, the marketplace rents real hardware by the second, inference endpoints turn any supported model into an OpenAI-compatible HTTP API, and the Playground lets you send a prompt to a live GPU without writing code. There are also SDKs. `pip install inferix-sdk` gives you the Python client and an `inferix` CLI; `npm i inferix` gives you the TypeScript client. Note the deliberate mismatch on the Python side: you install `inferix-sdk` and import `inferix`, because the shorter name on PyPI belongs to an unrelated project. ## What we will write about Practical things, mostly. How to pick a quantization. What it actually costs to serve a model once you account for idle time. How to tell whether a bad RAG answer is a retrieval problem or a generation problem. The kind of detail that only shows up once you have run something in production. We will also write about limitations. A post explaining a feature is more useful when it says where the feature stops being the right tool, and we would rather you learn that here than three hours into an integration. ## Where to start If you want to try something immediately, the Playground needs no setup. If you have a model in mind, the inference endpoints page will take you from a repository to a live URL. If you have a training job and no hardware, the marketplace is the shortest path. More posts follow. If there is something you want explained properly, tell us what it is.

Welcome to the Inferix community blog