Paste your existing resume into the first box, paste the job description you are applying for into the second, and get back a resume rewritten to maximize relevance for that specific role. The tool produces three things: an ATS keyword coverage table that shows which terms from the job description already appear in your resume and which are missing, a fully tailored resume with reordered bullets and tightened language, and a list of suggested manual changes the AI cannot make for you. The rewrite is strict about facts. It will not add experience, skills, or accomplishments you did not include in your source resume. It will reorder, rephrase, and surface what was already there.
A generic resume sent to twenty different job postings underperforms a tailored resume sent to ten of them. Recruiters and ATS systems both filter on keyword and phrase alignment with the specific role. A backend engineer resume that mentions Python, PostgreSQL, and AWS will rank below an otherwise identical resume that mentions Python, PostgreSQL, AWS, Kubernetes, and Terraform if the job posting calls out those last two. The skills are honest, the rewording is honest, but the alignment is what moves the resume from the silent pile to the shortlist. Tailoring used to mean rewriting from scratch for each role and most candidates skipped it. The tool collapses that work into a paste and a click while keeping you honest about what you actually have.
The first thing the tool produces is a coverage table. It pulls 8 to 12 important keywords from the job description and marks each as already present in your resume or not. The missing ones get a short recommendation. If the JD says "Kubernetes" and your resume says "container orchestration", the table flags this as a phrasing mismatch where you have the skill but the keyword does not match. If the JD says "5 years of Rust" and you have no Rust on your resume, the table flags it as a genuine gap and the rewrite will not invent Rust experience for you. This honest gap report is the most useful part of the output. It tells you what to talk about in your cover letter, what to study before the interview, and which roles are not actually a fit even if the title sounded right.
The tailored resume keeps your real experience intact and rewrites the surface. Bullets get reordered so JD-aligned achievements appear first under each role. Strong action verbs replace weak ones. Quantified outcomes get surfaced where the original implies them ("led a team" becomes "led a team of 6 engineers" if the original mentioned the team size elsewhere). Skills sections get reordered to match JD priorities. The structure of your resume stays the same, summary, experience, education, skills, so the output is recognizably your document, not an AI-generated replacement.
The suggested-changes section is where the tool admits its limits. It will recommend things like "add a metric to the bullet about the database migration if you remember how much latency dropped", "consider removing the 2014 internship which is no longer relevant to senior roles", "the JD asks for experience with Postgres, your resume says MySQL, mention if you have both". These are decisions you have to make, the AI cannot make them for you because they require knowledge of your actual experience that the resume text does not capture. Treat the suggested-changes list as a second pass, a checklist of things to verify, add, or remove before you send the resume.
Copy the full text of your existing resume into the first box. You can paste from a .docx, Google Doc, or PDF. The tool processes plain text, formatting like bold or italics is not preserved (and does not need to be, the output is markdown which you can paste back into your editor of choice).
Copy the full job posting into the second box. Include the role title, requirements, responsibilities, and any "preferred qualifications" section. The richer the JD text, the better the keyword extraction.
The tool sends both inputs to Claude with strict instructions to preserve facts and surface relevance. Generation takes 20 to 30 seconds for typical resumes.
Look at the keywords flagged as missing. Decide which ones represent genuine skills you can honestly add and which represent real gaps. The gaps tell you where to focus your cover letter and interview prep.
The output is markdown. Paste it back into Google Docs or your resume builder, apply your formatting (font, spacing, header style), and review the suggested changes list one more time before sending.
Mid-level software engineer applying to a senior role at a startup
The candidate has five years of full-stack JavaScript experience but the JD emphasizes TypeScript, Kubernetes, and event-driven architecture. The coverage table flags Kubernetes as missing, TypeScript as present-but-not-emphasized, and event-driven architecture as a phrasing mismatch (the resume mentions "pub-sub patterns"). The rewrite surfaces TypeScript and pub-sub more prominently and the candidate decides to learn enough Kubernetes basics to honestly mention it in the cover letter.
Product manager returning to work after a career break
The candidate has eight years of PM experience but a two-year gap caring for an elderly parent. The JD is for a senior PM role. The rewrite reorders bullets to put recent (pre-break) leadership accomplishments first and reframes the skills section to emphasize transferable skills. The suggested-changes list recommends a one-line summary acknowledging the break, which the candidate writes themselves.
Designer transitioning from agency work to in-house
The candidate has six years at design agencies but the JD is for an in-house design lead role that emphasizes stakeholder management, design systems, and cross-functional partnership. The coverage table flags design systems and stakeholder management as missing keywords even though the candidate has done both at agencies. The rewrite surfaces concrete examples of both from the agency work, which the candidate then expands on with the suggested manual additions.
Recent bootcamp graduate applying to junior dev roles
The candidate has a strong portfolio but only three months of bootcamp experience on the resume. The JD asks for one year of professional experience. The tool is honest about the gap, does not invent experience, and recommends emphasizing the portfolio projects in the cover letter and possibly applying to internship-track roles where the experience requirement is more flexible.
💡 Run the tailor before every application, not after writing the cover letter
The coverage table tells you what to talk about in the cover letter. Running the tailor first gives you the keyword targets for the letter, the order of skills to emphasize, and the gaps to address explicitly. Reverse the order and the letter is generic.
💡 Be honest about which gaps you can credibly close
If the JD asks for five years of Rust and you have none, applying is probably a stretch and the tailor will not paper over it. If the JD asks for Kubernetes and you have used Docker Swarm in production, that is close enough to mention in the cover letter while you brush up on Kubernetes basics.
💡 Keep your master resume long, let the tailor cut it down
A master resume with every project, every metric, every responsibility gives the tailor more raw material to surface. Trim for length only in the final output. The more text you give the tool, the better it can reorder for relevance.
💡 Save the coverage tables across applications
Over time the coverage tables reveal which keywords keep showing up in JDs for the roles you want. Those are the skills worth investing in, the ones that will move your resume from the silent pile across many applications, not just one.
No. The system prompt strictly instructs Claude to preserve all factual content and never add skills, jobs, or accomplishments the candidate did not include in the source resume. If a JD asks for a skill the candidate does not have, the coverage table flags it as a gap rather than fabricating it.
Copy the text out of the PDF and paste it into the input box. The tool does not yet accept PDF uploads for resume input. For ATS-friendly resumes the text copies cleanly. For heavily designed resumes copy may need cleanup, in which case extracting text from your original .docx or Google Doc source is the better option.
Currently the output is markdown that you paste back into your preferred editor and format yourself. Direct .docx export is on the roadmap and will let you download a Word document with your existing formatting preserved.
The tool will not surface skills that are not honestly in your resume. If you have a closely related skill (Docker instead of Kubernetes, MySQL instead of Postgres), the suggested-changes list will recommend mentioning the related skill in your cover letter. That is the honest path, not adding the missing skill to your resume.
Run the tool separately for each job. Each tailoring is specific to one JD, the coverage report and bullet ordering only make sense in the context of a single role. Save each tailored output in a separate file labeled with the company name.
The combined character limit is 6,000 characters on the free tier (enough for a 2-page resume plus a typical JD) and 30,000 characters with credits. Most real-world combinations fit in the free tier. If you are truncated, paste your most relevant 1-2 pages of the resume.
No. Both inputs are sent to Anthropic Claude for the tailoring step only. We do not store the resume, the JD, or the output. Each session is independent. Use the tool for confidential job searches without worrying about data retention.
The suggestions are sensible defaults but not personalized to your career. Treat them as a checklist of things to consider, not directives. The AI does not know whether the 2014 internship is actually a stronger predictor of your fit than the 2022 contract role, only you do.
Jobscan focuses on ATS keyword density scoring and gives you a numeric match score. This tool focuses on the rewrite itself plus a more selective keyword report. Use both if you want, the outputs are complementary. The honest-rewrite principle here is the differentiator, Jobscan and similar tools sometimes recommend keyword stuffing that does not reflect reality.