Skip to content

Beta.applyNchill is in active testing — expect rough edges, and check what it sends before you rely on it.

six stages · résumé agent · on your machine

Tailor

Rewrites your résumé for the one role in front of it, inside constraints that stop it inventing a career.

How this stage flows

Your queue and your résumé master go to apply-service, which claims a job atomically and assembles the finished briefing. A lane of the résumé agent on your own machine claims that briefing, renders a PDF, measures it and re-fits it if it does not fill the page. The result is one tailored PDF per job, attached at Apply.

the server builds the briefing — the agent never rebuilds it

claimed by a lane on your machine

rendered, measured, re-fitted if it does not fill the page

The server decides what to say. The agent decides how it fits on a page. Neither does the other’s job.

Tailoring is the stage with the most obvious way to go wrong: an unconstrained rewrite produces a résumé that reads beautifully and describes somebody else. The constraints are therefore the feature, not a limitation of it.

What it may not do

  • Counts are frozen. It cannot add a role, a degree or a publication.
  • It may never shorten the record. Reordering and rephrasing only.
  • Education and publications are untouchable.
  • Numbers may only come from a verified list. It cannot invent a percentage.
  • Adjacent skills are a valve with a limit, not an invitation to claim the stack.

Why it runs on your machine

The agent renders a PDF, and page fill is decided by measuring the rendered result rather than by asking the model to be brief. That needs real fonts and a real renderer, which is a poor fit for a container and a natural fit for the laptop the résumé belongs to anyway.

Work is claimed atomically, and only for jobs already queued to be applied to. Tailoring ahead of the queue is how you end up writing résumés all night for jobs nobody applies to.

Found this useful? Add applyNchill as a preferred source on Google