Runs are stored in the database and listed under Runs. After a run finishes, index successful results to Qdrant for vector search. The form below calls POST /eval/runs with your dataset and pipeline config.
POST /eval/runs
Create a dataset under Data → Datasets and add log entries first.
This sets normalization in the run config. It does not read config/config.yaml automatically; use regex to verify the pipeline, or model when Ollama is running.
normalization
config/config.yaml