CitedCV vs RecastCV

The short answer

Use RecastCV when you already have a CV you trust and want it rewritten for many job postings quickly: it takes a master CV plus a job URL and returns a tailored .docx in about half a minute, by its own figure. Use CitedCV when you are building the CV itself, want to be asked what is missing, and want the source of every line shown and checked before the file ships. Both say they will not invent experience; they differ in what they start from and what they show you afterwards.

Both are grounded, zero-hallucination tools: neither is meant to add a metric, a title or a project you did not supply, which separates them from ChatGPT, Claude or Gemini. This page compares two kinds of grounding.

A note on the word: in the UK, Europe and Turkey the document is a CV (özgeçmiş in Turkish); in the US it is a résumé. This page uses CV for both.

What RecastCV is

RecastCV describes itself as an AI CV-tailoring tool. You upload a master CV as PDF, DOCX or pasted text; it parses the file once and stores it as your source of truth, and you can add a project library of outcomes and metrics. For each application you paste a job description or its URL. RecastCV extracts the requirements and keywords, rewrites every bullet of the master CV in the language of the posting, shows a live ATS match score and exports a .docx you can edit inline first. Its about page describes a native iOS app that shares the same account and credits.

What it does well

It is fast: its feature page says most users get a tailored, ATS-ready CV in under 30 seconds. It is built for volume: one credit buys one tailored CV, credits never expire, it sells one-time credit packs rather than a plan, and a metric recorded once in the project library serves every rewrite. It preserves your voice and the structure of your CV and handles UK CV and US resume conventions. Around the CV it offers cover letters, interview prep, an application tracker, a free ATS checker and the iOS app.

Where it falls short

You need an existing CV; its own comparison page says so. If the master CV is thin or vague, the rewrite is bounded by it, and its grounding article admits a grounded rewrite can sound less impressive than an unconstrained one. There is no template library and no visual design; the native export is .docx, with PDF via print. Its pages say every bullet maps back to the master CV, but they do not describe showing that source next to each line, or a separate verifier after generation. The site is in English, and the conventions it names are the UK and the US.

What CitedCV is

CitedCV is an AI CV builder that writes only from facts you provide: a form, a PDF, DOCX or TXT file read in your browser, or your answers to interview questions. The unit is the fact, not the document. When something is missing it asks rather than fills. The writing model (MiniMax M3) receives the confirmed facts and nothing else, and every bullet carries the ids of the facts it was written from.

What it does well

Provenance on every line: open a bullet and you see the facts behind it. A deterministic verifier checks every number, date and named entity against your facts, and a language-model entailment check runs on top; a line that fails is rewritten once, then dropped, never shipped. A filler linter scores each CV for human-likeness. Templates are single-column with standard headings, exported as PDF and DOCX and re-parsed to confirm the file reads back correctly. Regional presets cover the US, UK, EU, Turkey and Germany, and the interface is in English and Turkish.

Where it falls short

It is slower on the first build, because it asks questions before it writes and verifies after; expect minutes, not seconds. It cannot write what you did not give it, so a thin fact set makes a thin CV. It costs credits: five are free at signup, a full build costs ten, and packs start at 30 credits for $9 (see /en/pricing). There are eight single-column templates, three free and five premium at 5 credits each. No cover letters yet, no interview prep, no application tracker and no mobile app; RecastCV wins on all of those today.

Side by side

How each tool says it refuses to invent

This is the question that brought most readers here, so it gets the vendors' own words rather than a summary.

RecastCV, in its own words

Its homepage FAQ answers "Does RecastCV make things up?" with a no: it rewrites strictly from the master CV and project library and never invents experience, titles or outcomes. Its feature page adds that a required skill absent from your CV is flagged as a gap rather than fabricated. Its grounding article describes three components: the master CV as source of truth, the project library, and rewrite rules that forbid metrics, technologies and job titles not in the source record; a draft that breaks a rule is flagged and regenerated. The same article offers five prompts for testing any tool, including a blank test in which a tool given only a posting should refuse to write.

The pages do not describe a second, independent check after generation, or a source shown to the user per line. The guarantee rests on the rewrite rules and on the input being your own CV.

CitedCV

CitedCV works from facts, each with an id, never from a document as a unit of truth. The writing model is called with the confirmed facts only. A deterministic verifier then checks every number, date and named entity in every bullet against those facts, and an entailment check confirms each bullet is supported as a whole. A bullet that fails is rewritten once and, if it fails again, dropped. The fact ids stay attached, so you can repeat the check.

When RecastCV is the better choice

• You have a CV you trust and are applying to many postings; one credit per tailoring makes the tenth application as easy as the first.
• You need a result in seconds, or on your phone between meetings.
• You want cover letters, interview prep and an application tracker drawn from the same master CV.
• You want to keep the structure and voice of the CV you already wrote rather than move to a template.
• You apply mainly in the UK or the US and want those conventions handled for you.

When CitedCV is the better choice

• You do not have a CV yet, or the one you have was written by an AI and you do not trust its numbers.
• You would rather be asked what is missing than have the gap flagged and left.
• You want to see the fact behind each line and have a verifier drop what is unsupported.
• You need a Turkish interface, or conventions for the EU, Turkey or Germany.
• You work in a field (finance, healthcare, public sector, law) where a line you cannot source can end an interview.

Honest verdict

If you already have a good CV and are applying widely, RecastCV is the better buy today. It is faster, it is built for volume, and its stated no-fabrication rule covers the step where most AI tools go wrong, the rewrite. Its extras are things CitedCV does not have.

If you are building the CV, or you want proof per line rather than a promise per document, CitedCV is the better choice. Its verifier and fact ids are the difference between being told a line is grounded and being able to check it.

Many readers will do well with both: build once in CitedCV, then upload that file to RecastCV as the master CV and tailor from there. A cited CV is a good master CV.

Using both: build in CitedCV, tailor in RecastCV

1. In CitedCV, enter your facts or upload your old CV, answer the questions it asks about gaps, and build. Anything dropped tells you where your facts were unclear.
2. Export the DOCX. This is now a master CV in which every line traces to something you confirmed.
3. Upload it to RecastCV as the master CV, add your documented metrics to its project library, and tailor per posting.

How to check either tool's output for invented claims

Both vendors say they do not invent. Check anyway; it takes ten minutes. RecastCV's grounding article proposes five prompts for any AI tool, and the list below folds them in.

1. List every number in the output and write its source next to it: the fact ids in CitedCV, the master CV and project library in RecastCV.
2. List every technology, tool and certification in the output that was not in what you gave the tool. The list should be empty.
3. Map every project named in the output to the place you described it. If one has no source, delete it.
4. Check for scope inflation: "contributed to" becoming "led", "mentored two people" becoming "managed a team". Your contract or LinkedIn profile is the reference.
5. Run the blank test: give the tool a posting and nothing else. A grounded tool refuses or asks; a tool that produces a full CV for a candidate who does not exist is not grounded.
6. Read the result aloud. A sentence you would not say is a sentence you cannot defend in an interview.

Questions people ask

Sources

1. RecastCV llms.txt: product summary written for AI crawlersrecastcv.com/llms.txt
2. RecastCV homepage: how it works and the FAQ "Does RecastCV make things up?"recastcv.com/
3. RecastCV pricing: credit packs, free credits, refund policyrecastcv.com/pricing
4. RecastCV feature page: tailor your CV to any job description in 30 secondsrecastcv.com/features/tailor-cv-to-job-description
5. RecastCV blog: why your AI resume builder is lying about you, grounding explainedrecastcv.com/blog/grounded-ai-resume-no-hallucination
6. RecastCV vs Rezi: build from scratch versus tailor the one you already haverecastcv.com/compare/recastcv-vs-rezi
7. About RecastCV: the grounded tailoring thesis and the iOS apprecastcv.com/about
8. Security at RecastCV: storage region and AI processingrecastcv.com/security
9. A survey on hallucination in large language models (arXiv 2311.05232, 2023)arxiv.org/abs/2311.05232
10. RoleWorth: best AI resume builder 2026, an honest comparisonroleworth.com/guides/best-ai-resume-builder-2026