uvero.meuvero.me

CV job match

Read your CV the way a recruiter reads the advert

The advert comes back colour-coded: shown in their words, shown in different words, only claimed, or absent. It is a working surface rather than a verdict — the requirement list is yours to correct, and the measurement moves with it.

Start free

Free to start. Have the advert open when you begin — the CV is built to answer it.

How it works

  1. Put the advert on the desk first

    Everything here reads from their side, so it starts with their words. Paste the full advert text when you create the CV. Creating it sends that text to a third-party AI provider on several separate passes — reading the role, extracting the requirements, and, when a first draft is composed for you, writing and reviewing it. What the score uses is the requirement pass: the phrases the advert names in its own wording, and how heavily it leans on each area. That reading is stored, and everything after it is arithmetic.

  2. See what they asked for, listed

    This is the shortlist they are working from, made legible. One list, ordered so the requirements you answer well sit above the ones you do not. Each row marks whether your CV shows it, shows it in different words, only claims it in a summary, or says nothing.

  3. See their words, coloured by your answers

    Now read their advert back, coloured by your own CV. The text is tinted in four tones by what you evidence. Wording belonging to no tracked requirement stays plain, so the colour describes the scored vocabulary, not the whole page.

  4. Disagree, fix, watch it move

    You are not obliged to accept the reading. Highlight a phrase it missed to start tracking it; remove one that does not apply with the × on its row. Then edit the CV — your own wording, or an AI rewrite of a single line. The score recalculates on every save.

What the four colours mean

A reader sees the page, so the page is what is compared — not the career record behind it. Whatever this version leaves off the page, because you switched it off or because the template had no room for it, sits outside the measurement. That is what makes tailoring measurable in the first place: two CVs cut differently from one record score differently.

Where a phrase sits decides what it is worth

Each bullet is weighed on its own, and position sets the value: a bullet carrying a concrete result counts most, and a requirement asserted only in your tagline or introduction counts for nothing at all. That bottom rung is the one worth knowing here, because it is the one that surprises people — a summary can claim anything, and one written by a model claims it fluently. The full ladder, with its multipliers, is written out on the scoring-method page.

Recency is folded in too: evidence more than about five years old carries reduced weight, while undated content is not penalised — unknown is not the same as stale. Depth is inferred from the roles a skill appears in, and that inference is yours to override: link a skill to the roles you actually used it in under My Career and the check reads those instead of guessing.

Two controls that make it yours, not the machine's

An advert is written by a person in a hurry, and one reading pass will sometimes miss the sentence the whole role turns on — which is exactly the sentence a human screener would fix on. So the requirement list is editable, and both edits reach the measurement itself rather than only the display.

Highlight a phrase to track it

Highlight wording in the advert — four characters or more, so a short token like "SQL" or "AWS" falls below the floor and produces nothing — and a "Track as requirement" control appears. The phrase joins the scored set immediately: it shows up in the rows, it tints where it occurs, and it moves the number. Added phrases are scored as hard skills and never as critical ones, so a manual addition informs the score without hijacking it.

Dismiss one that does not apply

Every row carries an ×. An advert mentioning a driving licence for a role you will never drive in should not sit there costing you points, so dismissing it drops that requirement from the score, the rows and the tint at once.

Both edits are stored apart from the machine reading, which is what makes them durable: if the advert is ever read again, a fresh analysis cannot resurrect a dismissed requirement or drop one you added by hand.

What it puts in front of you

A bare percentage is not advice, and it is not what a reader would tell you. So the number arrives with its parts: four areas — hard skills, context and recency, seniority fit, soft skills or domain — each with a value and the weight this advert placed on it.

Three of those four are fed by the advert's requirements, and an area it never raised scores full marks rather than zero, so you are not marked down for the shape of the advert. Context and recency is the exception, and worth understanding: no requirement feeds it. It scores your answers rather than the questions — how current and how deep the evidence behind them is — and it carries between a fifth and a third of the total. It reaches 100 only when the advert named no requirements at all, and 0 when the advert asked and your CV evidenced none of it.

  • Said, but not shown: requirements found only in your tagline or introduction, so you can move each into a bullet where it counts.
  • Answered in different words: the requirement's row carries a ≈ mark, and hovering it names the phrase that answered — coverage a literal keyword reader would miss.
  • Flags: each deduction that fired, named in plain language, with the total they cost and the score before they applied, so nothing is quietly subtracted.
  • What to improve: the reading's notes, beside the role title and seniority it took from the advert.

Those deductions measure concentration, not volume: one term saturating a section, one line carrying most of the advert's vocabulary, a year-long break in recent history. A specialist writing repeatedly about their specialism is not keyword stuffing, and concurrent roles are merged before any gap is counted. What each one costs is set out on the scoring-method page.

What the view does not show

It measures one thing, and it is not their opinion of you: how well this CV evidences the requirements this advert named, weighted as this advert weighted them. Two scores from two adverts came out of two different measurements, so lining them up as a ranking is the clearest way to misuse the feature. The comparison the method does support is the one this page is built around — the same CV against the same advert, before and after a round of edits.

And it is not a stand-in for the reader. It predicts nothing about an employer's decision: Companies run different applicant tracking systems, configured differently, and a shortlist is still drawn up by a person. What the comparison is good at is narrower and worth having: noticing that the capability an advert asks for three times appears nowhere in your first ten lines.

Questions people ask

Does every application need its own CV?
Every advert you want a match for, yes — the advert is attached when the CV is created, and the comparison belongs to that pairing. But it is selection rather than retyping: those CVs draw on one career record, so a bullet sharpened for one application is there for the next.
A CV started from a template has no advert attached, and so no match score.
The reading missed a requirement the advert obviously cares about.
Highlight the phrase in the advert and track it. It joins the scored requirement set straight away, so the rows, the tint and the number all account for it from then on. Twenty tracked phrases per CV is the ceiling — more than any real advert produces.
I added the missing keyword to my summary and the score did not move.
That is intended. A requirement mentioned only in a tagline or introduction scores zero, because asserting a skill costs nothing and evidencing it does not. Those cases are listed under "said, but not shown" so you can move each into a bullet — ideally one with a number in it.
Can I change the advert on a CV after I have created it?
No. The advert text is stored when the CV is created and there is no way to edit it afterwards, so if the posting changed or you pasted the wrong thing, create a new CV from the corrected text. What you can edit is the requirement list the reading produced.
There is a "Run the ATS check" button, but only where a CV has not been analysed against its advert yet — the one scoring step that calls a model on your instruction, and rate-limited accordingly. Once a score exists, the panel offers no re-read.
What gets sent to an AI, and when?
Creating the CV does. The advert text goes to a third-party AI provider on several separate passes within that one run — reading the role, extracting the requirements, and, when the first draft is composed for you, writing it and reviewing what was written. The privacy notice names the providers and the purpose, and is the authoritative document on how that data is handled.
Recalculating the score never does. Every recalculation is arithmetic over the stored reading, which is why saving an edit updates the number immediately and why the same CV against the same advert always gives the same result. Two later things do call a model, and both are things you ask for explicitly: the ATS check on a CV that has no score yet, and an AI rewrite of a single line.