Keyword match breakdown
Matched and missing keywords named exactly. No opaque score, no black-box percent.
Read onUpload one master resume during onboarding. Every job you track after that gets its own tailored version. The AI rewrites your bullets to mirror the posting. Same nouns, same phrasing, same order. A deterministic validator flags any role, education, or certification dropouts before the output reaches you. Nothing fabricated, ever.
4 steps. Each one runs in seconds. Scroll in order.
Onboarding asks for your full work history as one document. The AI also grades the master so you fix the foundation before any tailoring runs.
The parser extracts skills, keywords, and experience level. Those fields feed the tailor prompt so the rewrite is grounded in the exact posting.
The AI returns a structured shape (experience, skills, education, certifications). The server renders plain text and HTML on every tier, PDF watermark-free on every tier, DOCX on Pro Weekly, Pro, Pro Quarterly, and Founders. Every required skill from the JD either appears because it was already true in your master, or gets flagged as a gap.
Each regeneration costs one tailor credit. Pro accounts have unlimited monthly tailors within a 3-per-job lifetime regen cap that prevents reroll loops. Each version is saved so you can compare two rewrites of the same job.
One real example, same input reshaped two ways.
Worked on the web team building features in JavaScript frameworks. Shipped UI and worked with product and designers on various projects.
Partnered with product and design to ship 14 React and Next.js components into production, writing Jest unit tests and Playwright integration tests alongside each release. Held the team to a Core Web Vitals LCP budget of 2.5s.
You should be able to tell how a feature works before you trust it with your job search.
Tailoring is where model quality shows up. Pro accounts run on a stronger model for noticeably sharper output. Free accounts get three tailors per month on a faster, lighter model.
The system prompt is explicit: reshape and reframe what is already in the master resume. Never invent experience. Never add skills that are not in the master. If the JD asks for Kafka and the master does not mention Kafka, the output does not say Kafka.
No tables, no columns, no graphics, no header-or-footer placement for contact info, left-aligned, standard section names (Experience, Education, Skills). Plain text and structured JSON behind the scenes so format swaps do not require a new AI call.
After the AI returns, a non-AI validator checks every dated role in your master is present in the output, that education and certifications were not dropped, that the experience count matches. Then every "matched JD keyword" the AI claims gets re-grepped against your master resume. Anything the master cannot back up is dropped and logged. The matched count you see is the count that is actually grounded in your real history.
Three dimensions run side by side after a successful tailor: keyword coverage (weighted 50%), ATS format score (30%), and skill evidence in bullets (20%). The composite tells you the overall fit; the weakest dimension is marked "fix this first" so you know where the highest leverage is.
Honest limits read as trust signals. Hiding them does the opposite.
If your question isn't here, email support and you'll hear back from the founder.
No. The system is built around that guardrail. The prompt forbids adding skills that are not in your master resume, and the validation pass flags any mismatch. You get a list of real gaps, not a padded resume.
Around 10 to 15 seconds on the Pro model, closer to 7 seconds on the Free model. We show a live status message during generation so you know it is still running.
Yes, up to 3 times per job (lifetime cap on every tier, which prevents reroll loops). Each regeneration saves a new version so you can compare side by side. Free accounts have three tailors per month total (shared with cover letter, rejection analysis, and builder enhancements). Pro accounts have unlimited monthly tailors within the same 3-per-job cap.
Plain text and HTML are free on every tier. PDF export is free on every tier and watermark-free. .docx export is Pro Weekly, Pro, Pro Quarterly, and Founders only. We gate the format itself because .docx watermarks are trivially removable in Word, so watermark detection alone is not enough. Free accounts can also copy the plain-text version for manual paste into any resume template.
Yes. Resume builders give you a template and let you type in your history. This tailors a resume you already have to a specific job description, bullet by bullet. The master resume is the input, the tailored version is the output. We do ship a builder too (Pro+ feature), but the tailor and builder are separate flows.
A non-AI check that runs after generation. It counts dated roles in your master (anchored by year ranges like 2019-2023 or 2020-Present), compares to the role count in the tailored output, and flags if the output is short by two or more roles. It does the same for Education and Certifications sections. If the AI collapsed your "Earlier Experience" or dropped a cert, you see a warning before the resume is finalized, not a silently shortened version.
Every feature writes to the same record. These three are the closest neighbours.
Matched and missing keywords named exactly. No opaque score, no black-box percent.
Read onMark a job as denied. Get an AI read on the likely gap, what the resume missed, and what to do differently next time.
Read onDrop a raw job posting. Get structured company, role, skills, and ATS keywords back in under two seconds.
Read onCreate a free account in under a minute. First job tracked, first tailored resume, and first keyword breakdown all happen inside the onboarding flow.
No credit card · 2-minute setup · Cancel anytime