Reference
Exact lookup material and how things work: the platform, the catalogues, the file that describes your apps, and Thinkube Tandem.
Most days you work through the agent. Come here for the exact meaning of a field, a component, an operation or a word on the screen, and for how the parts behind it work.
The platform
-
Thinkube Control: the control panel’s pages, and the operations the agent calls.
-
LLM serving and models: model states, loading, sizing, routing, and calling the models.
-
Notebook execution over MCP: starting notebook servers and running notebooks from the agent.
-
What the install changes on your machines: every package, setting and file the installer puts on each machine.
Catalogues
-
Templates catalog: every template, what it deploys, and what it needs.
-
Components catalog: every component, what provides it, and whether it stays running.
Your apps
Every app on Thinkube is described by thinkube.yaml, one short file at the root of its repository. It says only what you know about your app: its containers, how each is built, the port it listens on, where requests go, and the services it needs, such as a database. At deploy time the platform generates everything Kubernetes needs from it. Thinkube GitOps explains what it replaces.
-
thinkube.yaml: every field, with examples.
-
What is in the web app template: what each file of the template is for.
-
How configuration reaches your app: the environment variables your app receives at deploy, and where each comes from.
Thinkube Tandem
-
The words on the page: every line, button and state word on Tandem’s pages.
-
Configuration and settings: Tandem’s commands and settings in Thinkube IDE.
-
The store: what Tandem records about your work, and where it is kept.
-
ODC (Orthogonal Defect Classification): the defect ledger for work done with an AI: what Tandem records about its runs to improve the process, and why.
-
The defect ledger, reference: the fields of a defect row.