Test Questions
Ask your knowledge a question and see exactly what it found — matches, sources and a verdict, never a polished answer. Save the questions that matter and run them again after every change.
At a glance
- Ask any collection a question, with your expected answer beside it
- Saved questions run one by one under Run all, counted as it goes
- Evidence, not prose — matches, sources and their percentages
- A verdict on every run — met, answered, missed or failed

Ask and Saved
One column holds two panels. Ask takes a collection, a question and the answer you expect, with Ask and Save side by side. Saved lists every kept question as a card — its text, a verdict chip or Not run yet, the expectation beneath, and Run, Edit and Remove; Edit opens a dialog, so the list never jumps.
The Result panel
The other column shows what the last run found: the question, what has changed since it last ran when anything has, the verdict badge, the best-match percentage, the sources used — each with its own match score and link — and the latency. There is no generated answer to admire; the panel lays out the evidence and leaves the judging to you.
- AskPick the collection, write the question and what you expect back, then Ask; Save keeps it for next time.
- Run the saved setRun all works through the saved list question by question, counting as it goes; each card can also Run, Edit or Remove on its own.
- Read the resultThe Result panel shows the verdict, the best-match percentage, every source used with its own match score, and how long the answer took.
Good to know — Run all counts as it works — 12 of 40 — so a long list always shows where it stands.
