What has changed on this site and when. Bug fixes are listed alongside new features, including the ones that reflect badly on us — a changelog that only lists launches is marketing, not a record.
A new tool that compares a job posting against your resume and reports which required terms are missing, which appear only as an acronym, and which you have over-used. It weights terms found under a "Requirements" heading above those under "Preferred", and ignores posting boilerplate such as equal-opportunity statements.
Salary-to-hourly conversion across every pay frequency, job offer comparison on total compensation, pay raise with an inflation-adjusted real figure, FLSA-aware overtime, time card, PTO accrual, freelance rate, contract-vs-permanent break-even, commute cost, 401(k) employer match, and severance package value.
The STAR builder validates that an answer names a result, is written in the first person, and fits the 90-second spoken band, and tracks which behavioral themes you have no story for yet. The application tracker keeps your pipeline in your browser with CSV and JSON export.
The check that looks for quantified results was matching any digit sequence, so an area code, a ZIP code, and a graduation year all counted as "metrics". A resume with no measurable achievements at all could be told it made good use of numbers — the exact opposite of the advice the tool exists to give. It now ignores contact blocks and bare years, and shows you the specific snippets it counted so you can check its working.
The phone detection pattern allowed only one separator character between digits, so "(555) 123-4567" — which has two — was not recognized. Users with a correctly formatted number were told to add one. Detection now normalizes to digits first.
Several output variants spliced a verb phrase into a slot that expected a noun, producing lines like "Increased results by the company social media accounts". It has been rebuilt around the X-Y-Z formula, and every candidate line now passes a validator — past-tense verb first, a number present, under 30 words, no first person — before it is shown. Candidates that fail are discarded rather than displayed.
Each new tool was verified by re-deriving its arithmetic from scratch and attacking it with edge cases, rather than by confirming it looked right. That found real defects in most of them and they were fixed before release. Among them: the time card divided minutes by 60 before applying the pay rate, so exact half-cent amounts rounded down by a penny; the offer comparison labelled the offer with the lowest insurance premiums as the "highest" on that row; the PTO calculator reported forfeited hours for a balance that exactly equalled the rollover limit, because of floating-point residue; the overtime calculator used the wrong divisor for salaried non-exempt staff who work fewer hours than their salary covers; the STAR builder read "savage" as the outcome verb "saved"; the CSV export from the application tracker had no guard against spreadsheet formula injection; and the salary page got its own worked example wrong, stating 26 x $2,615.38 as exactly $68,000 when it is $67,999.88.
The job description matcher page claimed that "roughly a third of a typical posting is company blurb, benefits and legal boilerplate" — a figure with no source and no measurement behind it, four paragraphs after the same page explains why you should distrust exactly that kind of number. It has been removed.
Long-form pieces on what applicant tracking systems actually do, quantifying a resume bullet when you think you have no numbers, five fully worked STAR answers, word-for-word salary negotiation scripts, explaining an employment gap, resume length by career stage, how pay frequency works, and how to read a job posting before applying.
Published what every tool checks and every formula it uses, the standard content is held to and what we refuse to build, the current accessibility status including known gaps, and a public record of changes. The site also now has a named author rather than being published anonymously.
Every dynamically added field in the repeatable sections rendered its label and input as siblings with no association between them, so screen readers announced an unlabeled edit field. None of the fields across the experience, education, projects or certifications sections were correctly labelled. Each field now gets a unique id and a bound label.
The resume templates page stated that the builder produces a single layout, while the builder has shipped a three-way template picker driving three genuinely different PDF layouts. The page now shows each layout with a preview, explains who it suits, and links straight into the builder with that template selected.
The resume builder and cover letter generator rebuilt the entire document preview on each character typed. Rendering is now debounced, which removes the main responsiveness risk on the two most interaction-heavy pages.
All 25 profession cover-letter pages led with an introduction written about resumes. Each now has its own opening written for cover letters specifically.
Several answers told you to use your browser print dialog and choose "Save as PDF". The builder has generated the PDF directly for some time — there is no dialog. The answers now describe what the tool actually does, including the three layout options.
The policy said users "may see a consent prompt". No such mechanism existed. The policy now describes the site as it actually is — no advertising cookies are set, because no ads are currently served — and states what will change before any ad is served.
A development placeholder reading "AD SLOT - disabled in dev" was being rendered into the production build on every page that reserves ad space. It is now shown only during local development.
Profession pages previously all linked to the same six roles, leaving 18 of 25 with no inbound links from their siblings. Related roles are now chosen from the same career category first, then rotated, so every profession is reachable.
The sitemap stamped all 63 URLs with the build timestamp, so each deploy asserted that the privacy policy and terms had just been modified. Those false freshness signals have been removed.
Published exactly what each tool checks and what it deliberately does not do, how advice is researched and corrected, and the current accessibility status including known gaps.
Most free tool sites give you no way to tell whether anyone is still maintaining them. A calculator using a tax figure from four years ago looks exactly like one updated last week. This page is the evidence, and it is the reason the pay calculators stamp every published rate with its year and link to the source.
It also records mistakes. Several tools on this site were shipped with real defects — a resume checker that counted phone number digits as career achievements, a bullet generator that produced ungrammatical sentences. Those are listed above with the date they were found and fixed. Publishing them is uncomfortable and it is the only way the editorial policy means anything.
Every tool is free, runs in your browser, and needs no account. Start with the resume builder or jump straight to a cover letter.