How it works
How dataset definitions become a hosted hypequery cloud release.
hypequery cloud is not live yet
Cloud deploys your dataset definitions, not your application code. A deploy has four steps:
- Publish. You choose which datasets cloud exposes and who can query them. In TypeScript this is
publishToCloud(...); in Python you write a deployment bundle directly. - Build. The CLI turns the publication into a verified deployment bundle. Datasets that published datasets relate to are included automatically.
- Release. The bundle is bound to a target project and environment as a release.
- Submit. Cloud accepts the release and starts serving it. The CLI prints the hosted REST and MCP URLs.
hypequery deploy runs all four steps for TypeScript projects. The individual deployment:* commands let you run
them separately, for example in CI or from Python.
What is deployed
A release contains a language-neutral deployment contract: your datasets, dimensions, measures, relationships, and the access policy for each published dataset. TypeScript and Python produce the same contract for equivalent definitions, so cloud serves both the same way.
By default the bundle also includes your project source files, so you can download them later with hypequery pull
or compare them with hypequery diff. Pass --no-source to leave them out.
Targets
hypequery login authorizes your machine and saves a default project and environment. Deploy commands use
that target unless you pass --project and --environment.