Documentation

StackYak runs AI infrastructure as one system: inference on a curated model catalog, compute on dedicated GPU nodes, and network between them. Take one layer or all three, composed into a Stack.

These docs are being rebuilt

The previous documentation described a product we do not sell, so it has been removed rather than corrected. The replacement is being written section by section, and each section is linked here as it lands.

Landed so far:

• API — the complete public inference surface, and what it deliberately is not.
• Changelog — customer-visible changes, from the October launch.

In progress: the platform overview, guides for running inference and dedicated compute, the rest of the API reference, the partner programs, and the release notes.

In the meantime

• The public API is https://api.stackyak.ai. It is OpenAI-compatible for chat completions and embeddings, so an existing OpenAI client works by changing the base URL and the key.
• Ask the Yak, in the product, answers questions against your real account data and will open a support ticket when you need a person.
• Anything wrong or missing here is treated as a product defect. Please report it.