User Guides

Jobs and Results

Understand statuses, reports, SHAP, and outputs.

Quickstart
Get from sign-in to a completed first job.
Choosing a Service
Match your data and goal to the right workflow.
Data Preparation
Prepare tabular datasets and media manifests for successful uploads.
Jobs and Results
Understand statuses, reports, SHAP, and outputs.
Billing and Credits
How charges, estimates, and refunds work.
Prediction Guide
Use trained jobs for manual and batch prediction.
Troubleshooting
Resolve common user-facing errors quickly.
FAQ
Short answers to the most common platform questions.
Reference
Handbook and Technical Reference

Open the handbook for the full product guide, then use the model-method appendix when you need mathematical details.

Open Handbook Open model-method appendix

Jobs and Results

Job statuses

  • Queued
  • Running
  • Completed
  • Failed

Where to monitor work

Use Workspace for the broad view across many jobs. Use the job detail page when you want to inspect one job closely.

Workspace is best for: - current status - recent jobs - live logs - artifact availability - quick navigation into prediction or manuscript actions

Job detail is best for: - one-job diagnostics - billing tied to that specific run - SHAP actions - manuscript generation and regeneration - report or prediction follow-up actions

What a completed job can include

  • service-appropriate metrics
  • model comparison artifacts
  • plots and explainability outputs
  • manuscript and executive brief outputs when generated
  • prediction-ready artifacts for supported services

How to read a completed result

Start with the best model

Confirm: - which model was selected - which metric drove that selection - whether the metric family matches the service type

Then review comparison outputs

Look for: - whether other models were close in performance - whether confidence intervals overlap materially - whether the best model is also operationally reasonable

Then review explainability

If SHAP or feature importance is available, use it to answer: - which variables matter most - whether the top signals make domain sense - whether any suspicious leakage-like predictors appear

Then review the AI report outputs

The AI Report tab or focused reader is for: - reading the journal manuscript - reading the executive brief - checking corrected figures and tables - comparing versions - checking provenance and export readiness - moving toward publishing or sharing

The AI Report tab on the job detail page is where you launch generation or regeneration actions. It also embeds the focused report reader after a report exists, so normal review can happen inside the job detail page without opening a separate route. The standalone /llm-report/<job_id> page remains available as a focused reader.

How to read billing on a completed job

Treat billing as a sequence of actions, not one monolithic charge.

Common pieces include: - training charge - SHAP add-on charge - AI report generation charge - AI report regeneration charge

Some add-on actions reserve credits first and settle later. On the job detail page: - Reserved Charge means the action placed a provisional hold - Net Charge means the settled amount actually kept - Refund means credits were returned after settlement or failure

If an AI report completed successfully but settled to 0, that usually means the report ran through deterministic fallback and no live LLM provider usage was detected for that run.

Important detail about SHAP

SHAP can be a separate add-on action for some services. A completed training job does not always mean SHAP already ran.

When a completed job still needs attention

  • SHAP has not run yet
  • manuscript generation has not run yet
  • the best model metric looks inconsistent with the task
  • expected artifacts are missing
  • billing or add-on status looks incomplete

When to regenerate a manuscript

Regenerate when: - the objective changed - explainability artifacts became available after the original draft - provider quality improved - the saved draft used fallback content and you want a stronger live-provider run

Related guides

  • Billing and Credits
  • Prediction Guide
  • Troubleshooting