Live preview runs your current, unsaved AI Qualification settings on the scanner's latest 50 jobs and shows what each would get. It is free, it never sends bids, and it does not change anything on your scanner.
How to run a preview
Open the scanner and go to the AI Qualification tab.
Change any settings you want to test: model, Ideal Client Profile, Hard Filters, Minimum Match Score or weights. You do not need to save first.
Scroll to Live preview at the bottom and click Run preview.
Wait for the results. A full run usually takes 1 to 2 minutes (a little longer with GPT-6 Luna). Rows appear as they are scored.
Adjust and click Re-run as many times as you need, then save when you are happy.
Reading the results
Each row shows one job with its preview verdict:
Qualified: the job would be sent to the Auto Bidder
Disqualified: the job would be filtered out or skipped by a Hard Filter
Not scored: the job could not be scored
Click a row to see the reason, which Hard Filter fired (if any), the score compared to your threshold, and the decision the job got when it first came in.
The footer sums it up, for example "18 of 50 would qualify, was 26". This lets you compare your new settings with what actually happened.
Limits
10 previews per person and 30 per team per day. The count resets at midnight UTC.
When you reach the limit, Run preview is disabled until the next day.
Clicking Run preview again while the same preview is still running does not use another preview.
Each person sees only their own latest preview for a scanner, since it runs their own unsaved settings.
The scanner needs at least one past job to preview.
Good to know
A preview never bids, never charges credits and never updates what the scanner has learned.
Each Re-run uses your current freelancer profile, agency profile and scanner memory, so results can shift slightly between runs.
On a trial, you can preview only scanners you own, and previews stop when the trial expires.