AI & ChatGPT · 2026-08-17 · 13 min read

ChatGPT vs a CV tool: when a chat is enough

Honest comparison: what ChatGPT or Gemini can do, what they cannot (ATS PDF, templates, privacy), and when ClarCV at €7/month makes sense.

The question is not “ChatGPT or ClarCV” as if they were the same product. ChatGPT (and Gemini, and Claude) is a general-purpose chat. ClarCV is a flow to upload a CV, analyse it, and export a PDF with an ATS template. They complement each other. The failure is treating the chat as if you had already sent the file.

If you want an alternative to ChatGPT for your résumé, it is usually because the wording is generic, you cannot turn it into a serious PDF, or you do not want to paste your career into chatgpt.com. This guide draws the line: when chat is enough, when it stalls you, and what a CV tool with AI does differently. No “future of AI” theatre: sending applications.

Start free at clarcv.com if you already have a PDF or Word file. Pro is €7/month with a clear price. Consumer chat looks free and charges you in editing time, invented facts, and sometimes in data you should not have pasted.

What the chat handles well

• Drafting a summary and bullets when you supply facts and an inventory.
• ES↔EN localisation with a “no calques” prompt.
• A keyword list from a job ad and gaps versus your text.
• Action-verb brainstorming from weak tasks.
• A tone pass (less marketing, more evidence) if you insist on it.

You do not need to pay for ClarCV on day one for that. You need a good prompt and a human review. The pack is in AI prompts for CVs. The method is in how to make a CV with ChatGPT. If your only problem is “I do not know how to write this bullet”, the chat is enough.

What the chat does not do (even if you ask)

It does not produce an A4 PDF with selectable text and a template built for parsers. It returns markdown, lists, or a “layout” that becomes an image when copied into Canva. It does not check whether your email lives only in a graphic header. It does not know if the file is too heavy for a portal. It does not fit the same content into 18 professional templates.

It also does not see your file the way an ATS does. You paste text; the employer filter receives a PDF. That gap explains applications that “looked fine in the chat” and get no reply. A CV tool works on the document, not only on the paragraph.

• ChatGPT: writes. It does not parse your PDF or tell you which section breaks on extraction.
• Gemini: can read an attachment, with imperfect extraction. Still no ATS template export.
• Claude: often invents less if you forbid it. Still text, not a sendable file.
• ClarCV: you upload PDF or Word, get clarity and structure analysis, regenerate with ATS templates to PDF.

Privacy: chatgpt.com versus a business API

In consumer ChatGPT you paste the CV into a generalist product. History and training settings depend on your account. There is no CV-specific policy and no DPA with ClarCV as controller. For a document with phone, email, and ten years of work, that matters.

ClarCV does not paste your CV into chatgpt.com. It uses OpenAI’s API for your request (analysis or rewrite), with a DPA, without using the content to train general-purpose models, and without selling data or running targeted ads. It does store what the service needs (account, file, results) while the account is active. Detail in what happens to your CV when you use AI and at clarcv.com/privacy.

No option is “zero servers”. The choice is the framework: a generalist chat or a CV product with stated rules. If you only want a nicer sentence, anonymise contact details in the chat. If you will iterate and download, a product flow reduces friction and surprises.

A concrete example (same person, two outcomes)

Ana has a Canva PDF and asks ChatGPT to “improve my CV for this CSM role”. The chat returns a glossy summary with “reduced churn 18%” she never measured, and a skills list copied from the ad (Gainsight included). She pastes it back into Canva, exports, sends. The ATS mangles the left column; the recruiter who does open the file doubts Gainsight in the interview. The chat did not fully fail: the system did.

Same starting point with the combined flow: Ana pastes the ad and inventory, the prompt marks [MISSING] on churn, she writes “40 HubSpot accounts, response time 12 hours to 5”. She uploads the PDF to ClarCV, switches to a single-column template, downloads selectable text. The ATS finds HubSpot and CSM. The interview is about facts she can tell. That is the difference between a chat tool and a CV tool.

Real cost (not only the subscription price)

ChatGPT Plus or Gemini Advanced are not built for ATS. You pay (or use a free quota) for a general model. The hidden cost is rebuilding layout every time, catching inventions, and wondering whether the PDF parses. ClarCV Free covers a first pass; Pro at €7/month makes sense if you tailor the CV to several ads in the same month.

If you apply once a year, chat plus a clean template can be enough. If you are in an active search, the bottleneck is not “another paragraph”; it is speed with control: one base, variants per ad, a reliable file. Then the chat becomes a step, not the whole product.

How to use them together (the flow we recommend)

1) Fact inventory outside the chat. 2) Prompt with the ad for a text draft (ChatGPT, Gemini, or Claude). 3) Catch [MISSING], clichés, and figures you never supplied. 4) Upload the PDF or Word to ClarCV: analysis + ATS template + download. 5) For the next ad, do not start over: adjust summary and bullets, generate the file again.

That is not “giving up ChatGPT”. It is not asking a chat to do a document generator’s job. The recruiter does not see your Gemini thread. They see an attachment. That attachment must paste into Notepad, show dates next to roles, and not contradict LinkedIn.

Signs the chat is no longer enough

• The Canva PDF will not let you select text, or the portal rejects the file.
• You have sent the same generic wording to ten different ads.
• The model inserted tools or percentages you cannot defend.
• You do not want to keep pasting a full CV into a consumer chat.
• You need ES and EN from the same facts with a stable layout.

If two or more apply, try ClarCV with the file you already have. The product AI flags vague wording and structural friction; templates avoid jumping to a layout that breaks parsing. Start free. If the analysis helps, Pro is €7/month—not an opaque enterprise plan.