**TL;DR.** Three things went live today. Your scanner now pulls from new job boards in addition to the big ATS providers (Recruitee, Teamtailor, Remote OK, We Work Remotely). It also stops dropping matches on the floor due to silent bugs that were quietly costing real interviews. And there's a new panel on the workers dashboard that shows which sources are actually finding qualified jobs for you, so you don't have to wonder.
What was wrong before
Two days ago a single scanner run looked at 21 active members and queued exactly 3 jobs total. That's not a slow week — that's a broken pipeline. Three different things were eating qualified matches before they ever reached your queue:
1. A salary parser would occasionally read free-text out of a job description ("we process 30M records per hour") as if it were the salary, then multiply it by a 2080-hour year, and try to write a 60-billion-dollar number into the database. Postgres refused. The job got dropped. So did the 40 other scored matches for that member, because the failure cascaded.
2. When a query against Google's API timed out (which happens), the scanner just gave up on that query. No retry, no fallback. Some weeks that ate half the source URLs.
3. The "queue is full" check was firing for members whose queues were full of stale jobs. Even fresh matches couldn't get in.
All three are fixed. If you saw a thin queue last week, that's why. You should see the queue refill from a much wider pool now.
What sources we now cover
Before today: 90 hand-picked companies plus the big ATS platforms (Greenhouse, Lever, Ashby, Workday, SmartRecruiters).
Now also: Recruitee (~10K small and mid-market tech companies, EU-leaning), Teamtailor (~8K, similar shape), Remote OK (high-volume remote board, refreshed daily), and We Work Remotely's programming feed. Plus a quiet improvement that matters more over time: any time the scanner finds a new company through Google search and that company produces a qualified match for any member, we automatically add that company to the global pool so future scans skip the search step and go straight to the source. The list grows itself, scoped to what people are actually applying to.
The dashboard panel
If you're a coach with workers dashboard access, the Scanner card now has a "Per-source yield" drilldown. One row per source over the last 1, 7, or 30 days. You can finally answer "which sources are producing qualified jobs for my engineers vs my marketers" without grepping logs.
What we held back
Hacker News "Who's Hiring" and a few enterprise ATS providers (BambooHR, Personio) are on deck but data-gated. We want to see seven days of real numbers on the freshness sources first before we add more, so we ship what actually moves the needle and skip what doesn't.
Nothing to configure here. The new sources are already on for every active member. If your queue was thin, watch it over the next cron cycle (every 8 hours) — you should see fresh inventory.