AGAuto Subjective GraderMultimodal Evaluation WorkspaceRun Evaluation

Evaluation Command Center

Configure a batch, run the multimodal grading pipeline, and follow each connected stage live from OCR to final report generation.

Total Recorded Runs0
Active Queue StateIDLE
Last Completed RunNo run yet
Best Recent ScoreWaiting for data

Launch A New Run

Choose the files and execution mode. The backend will create a run ID and handle the pipeline in the background.

Live Workflow

Watch the connected pipeline blocks update in sequence while the job runs.

No active run selected.Create a new job to activate the live workflow diagram and stage monitor.

Recent Run History

Reopen earlier evaluations, compare engines and OCR modes, and keep your testing workflow traceable as the dataset grows.

Loading runs...The API is collecting previously saved evaluations.