-
The Application Kit is now a Prepare, Apply, Track hub
Each job's kit walks the application as three stages with one focused block at a time, and the Apply stage ends at a submit gate only you can click. RoleReady never submits for you.
-
A stage-aware rail on the job page
The side rail now shows only the cards that make sense for where the job actually is, and there is one clear place to tailor a resume instead of several.
-
Interview Cheat Sheets
Generate a one-page cheat sheet from your resume and the job description, then read it in the app, print it, or download the PDF for your second monitor.
-
Quick Actions
Type a sentence like 'add a contact at Stripe' and it happens. The AI Assistant became Quick Actions with a pinned sidebar entry, 10 free runs a day, and a review step before anything runs.
-
The Knowledge Bank, organized into four dimensions
Tabs are now Experience, Skills, Education, and Projects & more, and the readiness panel keeps suggesting next steps instead of disappearing at 100%.
-
Impact receipts on AI results
AI results now tell you when they were built from your Knowledge Bank, so filling it in visibly pays off.
-
Every quota, visible
All metered features now appear in the plan tables, the resume tailor shows its meter before you hit the limit, and the 40-job wall opens the plan dialog instead of a dead-end toast.
-
A tidier calendar
Filter chips for event type and status, a Schedule round action that opens the event dialog right where you are, and a faster first load.
-
An honest wait after capture
One shared extraction clock across the web app and the extension, plus a status strip on the job page that shows real progress while intelligence runs.
-
Greenhouse postings capture correctly
Job pages embedded from Greenhouse no longer produce empty captures or an eternal loading spinner.
-
A friendlier front door
New accounts get one welcome at a time, the flow stops re-asking what it already knows, sign-up lands on the app instead of a 404, and the password meter agrees with the submit button.
-
Per-task AI model and tier picks
Each AI task now uses the model and tier its live benchmarks justify: the best model per tier, and 13 tasks moved down a tier where quality held while cost and latency dropped.