FAQ
What service should I start with
Start with the data shape shown on the homepage, then match the goal.
Is cross-sectional the same as spreadsheet data
Yes. In the current UI, cross-sectional means the familiar one-row-per-record spreadsheet or CSV workflow.
Why do some URLs or logs still say tabular
The user-facing product has shifted to cross-sectional language, but some route names and internal service identifiers still use tabular.
Do jobs keep running after logout
Yes. Jobs should continue on the server even if the browser session expires.
Is SHAP always included
No. SHAP is a user-initiated add-on from the SHAP tab/API; training only prepares the saved model and artifacts it needs.
Why was I charged
Charges can come from training, SHAP, or AI report actions.
Why do I see a refund after a successful AI report
Because AI report billing is settled after generation. A report can succeed operationally through deterministic fallback while settling to 0 if no live LLM provider usage was detected, so the original reserved charge is refunded.
Why are there multiple billing rows for one job
One job can have several separate billing actions such as training, SHAP, AI report generation, and AI report regeneration. Some of those actions also reserve credits first and settle later.
What does Net Charge mean
It is the final settled cost after any provisional hold and refund are reconciled. It is the number you should use when deciding how much the user actually paid for that action.
What if a prediction page says the model failed to load
The trained artifact or loader may not match. Check the job detail page and troubleshooting guide.