Privacy

What we hold, and what we do with it.

This describes how WellParsed handles your information. It is written to match what the product actually does. Where something is not yet true, it says so.

Last updated 6 September 2026.

What we collect

• Your resume. The file you upload, and the text and role history read out of it.
• Job postings. The URL or the text you paste, and the requirements read out of it.
• Your answers. What you type or say in the interview, including voice transcripts.
• What we build from those. Your experience library, drafted claims and composed resumes.
• Your email address, when you create or sign in to an account, or if you leave it in the box on the landing page. You can check whether a public job link works before giving us an email address, but an account is required before career information is saved.
• Operational records. Which model handled which task, how long it took and what it cost, so the service can be metered and debugged.

Where your work is stored

On our servers, after you create or sign in to an account — not only on your device. The server reads the public job link in memory during the first compatibility check and does not retain the full URL or posting. Your browser temporarily remembers the public link so it can continue the step after you sign in. If a site cannot be read, we record only its hostname, the detected career-site platform and a failure category so we can decide which reader to build next. After account creation, uploaded files are held in object storage and everything else is held in our database.

Who else sees it

To do its job the product sends parts of your information to service providers. They process the request and return a result. Nothing is published, sold, or shared with employers, and nothing is posted anywhere on your behalf.

• Language model providers, reached through OpenRouter: job text, resume text and your answers, in order to read postings, draft claims and run the review pass.
• ElevenLabs, for voice interviews: your speech, processed in the cloud rather than on your device.
• Stripe, for payments. Card details go to Stripe directly and never reach us; we hold only the customer and subscription identifiers.
• Cloudflare R2, which stores your uploaded resume file.
• Inbound, which delivers sign-in emails.

Operators who have been granted access to the admin console can read account content — resumes, answers, transcripts, composed text and stored files — in order to run the service. That access is not used to publish, sell, or share your work with employers.

Errors are recorded in error tracking we run ourselves. It is configured to carry operational detail only — identifiers, counts and reasons — and never the content of your resume, answers or postings.

Model training

We do not train models on your information and we do not give it to anyone to train on. Every request we send on your behalf carries a no-retention instruction, so it is routed only to providers that do not store what we send them. That instruction is attached to every call rather than to the ones we remembered, and a provider that will not comply is refused rather than quietly used.

Voice

Voice is optional. The microphone is requested the first time you use it, never at launch. Audio is not stored anywhere by us. The transcript is, because your own words are the evidence behind every line the product writes.

Deleting your data

Settings has a Delete my data control for signed-in accounts. It removes your uploaded files from object storage and your rows from the database — resume, roles, library, answers, transcripts, claims, resumes and sessions — immediately rather than on a schedule, and the removal is verified rather than assumed. It cannot be undone. Records we must keep for tax and accounting, held by Stripe, are not ours to delete.

It also removes the funnel record and any captured email address described in Cookies and tracking below. The funnel record goes in two pieces, because that is how it is stored: the steps already recorded against you, and the steps left under any visitor identifier of yours that had not been attached to anyone yet — a visit from the same browser, before you ever opened the app, is the ordinary case. A step under that identifier that belongs to a different account is that account's to delete, not yours. The captured email follows the same visitor link, or matches directly if you later sign in with the address itself. Unlike the removals above, neither is individually verified: the funnel rows go by cascade and by visitor identifier, the captured email by a plain delete, with no check afterward.

How long we keep it

Until you delete it, using the Delete my data control above. There is no automatic expiry for career information: your experience library is the thing that makes the tenth job easier than the first, so it is kept until you say otherwise.

That control can only run against a session that exists. Two things never create one: a bare landing-page visit that never goes on to open the app, and an email address left there by someone who does the same — both are then kept indefinitely by default. The email address is the one part of that we can still act on: write to us with it and we will remove it by hand, see Your choices. A page view with no address attached gives us nothing to search for, so it has no removal path at all.

Cookies and tracking

No advertising trackers and no third-party analytics. Before account creation, the app stores a technical session identifier on your device so it can rate-limit link checks and connect the visit to an account if you continue. That session does not own a saved job, resume or interview answer.

We also measure our own product, on servers we run. The landing page and the app each store a visitor identifier in your browser — never a cookie — and, if a link brought you here, the short code from that link and when it first did, so we know which of our posts is worth writing more of. The app also keeps a memory of two specific events so opening the same result twice does not inflate the count: a finished resume you open again, and the paywall shown while a resume is still generating. Every other step has no such guard — reopening the app, adding another job, or landing on the pricing screen's paywall again each count as a new event.

We record which steps you reached — that you opened the app, added a job, uploaded a resume, reached the paywall, started checkout, and so on. Some of those records carry a small label describing how it happened: the kind of file you uploaded, how the job was added, which plan, which paywall screen you saw. Besides the step, that label and when it happened, each row also carries the visitor identifier, the campaign code if one brought you here, and the internal id of the technical session or account involved — anonymous during the link check, replaced by your account's own id, the same one on every device you sign in on, once you continue. It never contains the content of your resume, your answers or the postings you paste. That record is kept on our own servers and shared with nobody.

Before you have an account, the landing page also has a box to leave your email address in. That address is stored in its own table, together with the visitor identifier, the campaign code if one brought you here, and when you submitted it, so we can follow up and credit the post that earned it. Once you create or sign in to an account, Delete my data can reach an address tied to the same visitor identifier. Signing in later with the same address works too, matched directly rather than through the visitor identifier. The one address that link never reaches is one left by a browser that never opens the app at all — see How long we keep it, and write to hello@mail.mvp-labs.com to have that one removed by hand.

Your choices

• Check whether a public job link is readable without an account or email address.
• Use it without voice, and without uploading a resume.
• Delete your account and its career information at any time from Settings.
• Left an email address on the landing page and never opened the app? There is no session for Delete my data to run from — write to the address below and we will remove it by hand.
• Ask what we hold about you by writing to the address below.

Children

WellParsed is not intended for anyone under 16, and we do not knowingly collect their information.

Changes

If this changes in a way that affects what happens to your information, the date at the top changes with it.

Contact

Questions about any of this: hello@mail.mvp-labs.com.