AI Policy

Effective Date: September 4, 2026  |  Last Updated: September 4, 2026

Several parts of Revquix use a large language model. This page explains which ones, what we send, what we do not, and what you can and cannot rely on. It forms part of our Terms of Service and is incorporated into them by reference.

1. Where AI is used

• Career tools. Resume readiness reports, job-description matching, bullet rewriting, cover letters, outreach messages, interview questions, summaries and similar. The AI writes the explanations, the suggestions and the drafted text.
• The code assistant. Explaining an error, explaining code, fixing, optimising, reviewing, documenting, generating tests, writing a README, and explaining a whole project.
• Practice help. Explaining why a submission failed a particular test case, and related coaching on a coding problem.

Each of these is something you start. Nothing on Revquix sends your content to a model in the background.

2. Who processes it

Our AI provider is DeepSeek, listed with everything else on our Sub-processor list. Your Revquix account identifiers are never sent to them - a request carries the text being worked on, not who you are.

Two things people often assume are AI services running elsewhere are not: your code executes on our own servers in a sandbox, and the model behind semantic search runs on our own infrastructure.

3. What we redact, and what we do not

3.1 Resume and job-description text - redacted automatically

Before resume or job-description text leaves our servers, we replace your name, email address, phone number, postal address, date of birth and web links with placeholders, and replace each employer name with a neutral label such as “Company 1”. This happens on every such request, without you having to ask for it, and it is why the reports you get back talk about “Company 1” rather than your last employer.

3.2 Source code - not redacted

When you ask the code assistant about a file, that file is sent as you wrote it. We do not redact code, and we want to be direct about why: there is no reliable way to tell a password from an ordinary piece of text, and a filter that removed the wrong strings would silently give you answers about code you did not write.

So: do not leave passwords, API keys, tokens, connection strings or other people's personal data in a file before pressing Explain, Fix, Optimise, Review, Document, Generate tests, README or Explain project. The project-wide action sends the most, because it has to see the project to explain it.

If you have already done this, the fix is the same as it would be anywhere: rotate the secret. Removing it from the file afterwards does not un-send it.

4. What the AI does not do

Your scores are arithmetic, not a judgement call by an AI. Every dimension of a readiness score is measured in our own code, weighted as published, and added up - the contributions sum to the total, so you can check it yourself. The AI writes the explanations and the rewrites; it never produces a number and is never asked to re-judge a finding.

Nor does any AI on Revquix decide anything about you with a legal or similarly significant effect. It does not set prices, gate access, approve or decline a booking, or rank you against other people for anything that matters. There is no automated decision-making of that kind on the Platform.

5. Accuracy, and what you should not rely on

• AI output can be wrong. It can be confidently wrong, out of date, or subtly mistaken in a way that reads well. Read what it gives you before you use it.
• It is not professional advice. Nothing an AI feature produces is legal, financial, immigration, medical or employment advice, and it is not a substitute for a professional who knows your situation.
• It does not guarantee an outcome. A high readiness score is not a prediction that you will be shortlisted, and a generated cover letter is not a prediction that you will be hired.
• Code it writes is a suggestion. Review, test and understand a patch before you apply it. You remain responsible for the code in your project.
• You are responsible for what you submit.Do not submit other people's personal data, confidential material belonging to an employer, or content you have no right to share.

6. Training

Revquix does not train models on your content. We do not build or fine-tune models from your resumes, your code, your projects or your reports.

What our AI provider does with a request once it reaches them is governed by our contract with them and by their own terms, and we will state their position on this page once we can state it precisely rather than approximately. Until this paragraph says otherwise, treat anything you send to an AI feature the way you would treat anything sent to a third-party service: do not send secrets, and do not send other people's confidential material. Section 3.2 is the practical version of that advice.

7. Who owns what

• Your input stays yours. Sending your resume or your code to an AI feature does not give us any ownership of it.
• You may use the output freely - in your resume, your application, your project. We claim no ownership of it and no royalty on it.
• We cannot promise output is unique. A model given a similar prompt may produce similar text for someone else, and generated text is not something we can warrant as original or as protectable by copyright.
• Check before you present it as your own work. Where you are asked to certify that something is your own - an application form, an assessment, an academic submission - that is between you and whoever asked.

8. Your choices and your records

• Nothing is automatic. Every AI feature is one you press. Not using them is a complete opt-out, and the rest of the Platform works without them.
• We record which run you made and what it cost, along with the version of this notice that applied at the time, so that a question about a past report can be answered against the terms that actually governed it.
• Reports are kept for you, and then deleted. Uploaded files, extracted text and generated reports have retention periods set out in Privacy Policy §6.
• Personalisation is separate and switchable. What we infer about your interests, and how to turn it off, is Privacy Policy §2.11 and §7.3.

9. Using AI features acceptably

The rules on manipulating prompts, misrepresenting your experience with generated text, and reselling our AI features are in our Acceptable Use Policy §5.

10. If something goes wrong

If an AI feature produces something harmful, defamatory, or plainly wrong about you, tell us at [email protected] with the report or run in question. A person will look at it, and we can correct or remove a stored report.

11. Changes

We may update this page as our AI features and providers change - including, in particular, Section 6. We will revise the “Last Updated” date, and material changes will be notified in-product or by email.