AI résumé builders, and what a tool without one does instead
Several builders say on their own pages that a model writes or checks for you. This one calls none. Here is what each approach can and cannot do, with the vendors' own wording.
— no model, anywhere.
Almost every résumé product now offers a model that writes for you, and the offer is genuinely attractive: an empty page is the hardest part, and fluent text in four seconds solves it. This product does not offer it, and the reason is worth setting out rather than asserting.
Where a vendor is named below, the wording comes from their own public page, read on 12 September 2026, with the link.
What the vendors say
Kickresume's pricing page lists an AI writing feature and an AI Resume and Cover Letter Writer under its paid plans, and describes the tool as letting AI write your résumé for you. Enhancv's pricing page describes AI-assisted tools including résumé checking, bullet point generation and job tailoring. Microsoft's résumé templates page offers Copilot to help polish, enhance or edit résumés using AI.
Those are three different offers, and only the first is really about authorship. Polishing a sentence you wrote and generating a bullet about work you have not described are not the same act.
The property that makes this awkward
A language model produces text that is plausible given what it has seen. It has no way to know which of your projects were real, and it will fill a thin section with something reasonable, because that is the shape of the task it was given.
On a résumé that produces a specific failure: a line that reads well, that you did not write, about work you cannot describe in detail when somebody asks. The document's job is to survive a conversation with a person who will ask about it.
What this product does instead
Three deterministic mechanisms, each of which can be inspected.
The reading. Specialist readers run over rules to pull a document into fields, and every string a reader emits is checked in code to be a real span of the source. A field that fails that check is dropped and counted rather than kept, which is described in dropped fields and why they are counted. A field it could not find is reported as not found, never filled in.
The score. The ATS checker is arithmetic over rules, and it prints the multiplication. A measured run on an invented candidate's résumé read 159 words, 20 lines and 4 recognised headings, and published 100 times 30 per cent plus 100 times 30 per cent plus 92 times 40 per cent, which is 97. Nothing in that sentence requires you to trust anything.
The letter. The cover letter builder selects sentences from the material you gave it and carries the line each one came from, so a claim can be walked back to its source. Where the evidence is thin the letter is shorter. Requirements from the posting that your evidence cannot support are listed as gaps and never asserted, which is the subject of why a letter only claims what the résumé shows.
Side by side
| A generative builder | Conxfolio | |
|---|---|---|
| Empty page | Filled in seconds | Still yours to fill |
| Source of the sentences | The model | Your own material |
| Traceability | Depends on the product | Each proof paragraph carries its source line |
| Thin evidence | Can be padded | Produces fewer paragraphs |
| Unsupported requirement | Can be claimed | Listed as a gap |
| Scoring | Often a number | A number with its arithmetic printed |
| Cost to run | A bill per use | None, which is why it is free |
Where a model genuinely helps
Getting unstuck. If you cannot start, generated text is a prompt to react against, and reacting is easier than inventing. Rewriting your own sentence more concisely is also a reasonable use, because the content is already yours.
Nothing stops you doing that elsewhere and bringing the result here. Write in whatever you like, paste it into the import, and use this side for the structure, the measurement and the rendering.
The commercial consequence
Calling no model means there is no per-visitor bill, which is a large part of why everything here is free with nothing held back. A product that pays for inference on every draft has to meter somebody. That connection is rarely stated and it explains most of the paywalls in this category, as the page on making a résumé without paying sets out.
What neither approach can do
Neither predicts whether you will be filtered out, because neither has access to an employer's applicant tracking system. Any product claiming to beat one is claiming something it cannot know. The checker here says so in its own words on the source step: the score comes from the structure and evidence it can read, and it is not an employer's private system.
Reading a claim about AI on any builder's page
Three questions make the marketing legible. Does the model write sentences, or only rearrange sentences you wrote. Can it produce a claim about work you have not described. And does the product show you where each sentence came from, or ask you to check the output yourself.
Those distinctions matter because the risk is not the technology. It is a document you will be asked about in a room. A tool that generates and shows its sources is a different proposition from one that generates and hands you the result. The published free tiers of two design-led builders are quoted on the Enhancv comparison and the Kickresume comparison, including which AI features each puts behind a plan.
See what deterministic looks like
Import a résumé and read what the reading returned, field by field, beside the text it came from. Free, no account.
Open the résumé builderRead the scoring arithmeticQuestions
- Does Conxfolio use a language model
- No, nowhere in the product. The document readers, the ATS score and the cover letter engine are deterministic code, and the letter's own response reports that no model was used.
- Which builders say they use one
- Several say so on their own pages. Kickresume's pricing page lists AI writing feature and AI Resume and Cover Letter Writer under its paid plans and says the tool will let AI write your résumé. Enhancv's pricing page names AI features. Microsoft's résumé templates page offers Copilot to polish, enhance or edit résumés using AI.
- Is a model bad at this
- That is not the claim. A model is good at producing fluent language quickly. It is also capable of writing a sentence about a project you did not do, which is a bad property for a document an employer will ask you about.
- How does a letter get written without one
- By selection. Each evidence paragraph is a sentence from the material you gave it, carried with the line it came from. Where there is no evidence it returns fewer paragraphs, and requirements it cannot support come back as gaps.
- How is a résumé read without one
- By deterministic readers over rules. Every string a reader emits is checked in code to be a real span of your document, and one that fails is dropped and counted rather than kept.
- Does that mean the writing is worse
- It means the writing is yours. The tool arranges, measures and renders, and the sentences come from you.
Conxfolio is a free set of four career tools: a résumé builder with 37 rendered layouts, an ATS résumé checker that prints its own arithmetic, a cover letter builder that traces every proof paragraph back to the line it came from, and a portfolio builder with 20 authored designs. There is no account to create, nothing is held back for a paid plan, and no language model is used anywhere in the product, so the readers, the score and the letter are deterministic code you can check.