Results and files
Rating a run
Tell us whether a result was useful, and what happens to the rating you leave.
Under the last turn of a chat sits a single question: was this useful? Answering it takes one click, and it is the most direct way to tell us what is working. Nothing about your run changes when you answer, and you can change your answer later.
Where it appears
On the most recent turn of a chat, once that turn has finished, failed, or stopped part way. Older turns do not carry it; the intent is a quick reaction to the result you are looking at rather than a survey of everything you have ever run.
Both a thumbs up and a thumbs down are recorded. A thumbs up is a complete answer on its own; a thumbs down opens a short form so the rating has something actionable behind it.
What the form asks
| Field | What it is for |
|---|---|
| What went wrong | A list of the usual culprits: incorrect RTL, a missed requirement, verification that was too thin, a simulation or synthesis problem, a slow result, a poor explanation, an unexpected cost, or something about the interface. |
| Anything you want to add | Free text. One concrete sentence is worth more than a paragraph of adjectives. |
| Would you use this output? | As generated, with minor edits, with significant edits, or not at all. This is the single most useful signal on the form. |
| Detailed ratings | Optional scores for RTL quality, requirement correctness, testbench quality, verification usefulness and how useful the synthesis information was. |
| Share this run with our team | Off by default. Switching it on lets us open the run itself while we look into what you reported. |
Everything except the thumb itself is optional. Submitting with only a reason is fine.
The sharing checkbox
Ratings are useful; a rating we can reproduce is far more useful. That checkbox is the only thing that lets someone here open the chat you rated, and it is off unless you tick it.
- It applies to that run, not to your account or your other chats.
- It is not a share link, so nothing becomes reachable by anyone outside our team.
- If you would rather send it yourself, a share link on a support ticket does the same job and keeps you in control of it.
What happens to it
Ratings are read. They feed the decisions about which models sit behind which role, where verification needs to be stronger, and which parts of the interface keep tripping people up. What they are not is a support channel: nobody replies to a rating.
Tip
If you need an answer, file a support ticket instead, or as well. Rate the run to tell us it happened; open a ticket to get it fixed for you.
Changing your mind
Rate a run again and the new answer replaces the old one. That matters more than it sounds: a module that looked wrong until you read the testbench is worth re-rating, and a module you approved and then found a bug in is worth re-rating too.