Privacy Policy

Last updated: October 12, 2026

What we collect

Account data: your email and name from Google sign-in (via Supabase Auth).
Resume content: what you type into the editor. Without an account it stays in your browser's local storage; with an account it is also synced to our database.
Payment data: handled entirely by Paddle — we never see or store card numbers, only the subscription reference and its status.
Feedback: messages you submit through the feedback widget.

How we use it

To provide the editor, cloud sync, PDF export, and portfolio pages.
AI features (parsing, analysis, bullet writing, cover letters) send the relevant resume text to our AI provider (Groq) to generate the result. It is not used to train models.
Import from GitHub: when you paste a public repository link, we read selected public files and metadata from that repository through the public GitHub API: its description, primary language and topics, the list of its file names, its README, its project manifest files (such as package.json) and, if present, up to two documentation files (such as FEATURES.md, ARCHITECTURE.md or OVERVIEW.md). We do not read source code files unless you choose the optional code analysis described below. Optional code analysis (available only when it has been switched on for the service, and only for an import for which you tick a separate box): our server also reads up to 6 selected source files and up to 3 manifest files of that public repository from GitHub (GitHub receives those read requests). They are scanned for secrets, never run, and not stored. From them our server extracts short, standardised descriptions of what the code contains (such as the names of pages, API routes, data models and screens, and the services the code calls) and keeps them only in temporary server memory while that one request is handled. Up to 3 draft bullet points built from those descriptions are returned to your browser for you to review and edit; the source code and extracted descriptions are not sent to your browser, but the response lists paths of inspected README, manifest and documentation files in its evidence summary. The code-analysis result reports counts and which bullets came from code, without listing the additional source-code files it inspected. None of this code-derived information is sent to Groq or another external AI provider, and none of it is stored in our database, during the import; bullets you choose to add to a resume are saved and treated like the rest of your resume text (see below). A future version might send only the short descriptions (never code) to Groq, and it would ask for your consent again first. We try to remove anything that looks like a secret (this filtering is best effort and not a guarantee, so please do not import a repository whose public files contain secrets). For the standard public GitHub Import, when its AI drafting step is used, our server first reads the README and repository details locally and matches them against a fixed, reviewed list of project types, domains, features and technology names (technologies are included only when the README mentions them or they are the repository's primary language). Only the matching list entries are sent to Groq, each labelled with a number and its kind, so that it can write the draft description. This standard import step does not send the README text, repository or owner names, description, topics, file names or paths, source code, or anything derived from code, including technologies read from manifest files. Every sentence the AI writes is checked on our server against exactly the entries it was given, and anything they do not support is removed. The checked draft, together with short statements quoted by our server from the README and (only if you ticked the separate code-analysis box) the code-analysis bullets, is returned to your browser for you to review. When the README is too short for the AI step, or the AI service is not configured, the draft is built on our own server from the README statements, the file names and the manifest files, nothing is sent to Groq, and the import dialog tells you that no AI wrote it. We do not store the repository content in our database; to avoid repeated requests to GitHub, the public, filtered text of the files we read (README, manifest files and any documentation file) and a summary of the file layout (including the names, sizes and Git identifiers of the repository's source and manifest files, but not their contents) may be kept in temporary server memory for up to 10 minutes and is then discarded. Only the project you review and add is saved.
Optional GitHub Professional Writer: for a public, whole-repository import, you can separately choose an offline, source-cited wording candidate for review. This offline candidate does not send its public repository evidence to Groq. If the AI wording option is available, it requires a separate, explicit version-2 consent for that import. Only after you tick that AI box may our server send Groq up to 8 selected, screened sentences from the public repository's README, description or project documentation, plus technology names detected in dependency files or reported by GitHub as the repository's language, so Groq can suggest resume wording. Our screening drops sentences when it detects links, email addresses, file paths, code, apparent secrets or instruction-like text; this filtering is best effort. A sent sentence can still contain names mentioned in the public evidence, including the project's name or a person or company name. This optional request does not send whole files, private repositories, source code, your resume or account details, GitHub tokens or the repository link. Groq processes the information sent to it on its servers and may temporarily retain inference data under its own policy unless Zero Data Retention is enabled for our account. You review and choose any suggested lines before they become saved resume content.
Optional private-repository wording (separate consent): for a private repository that you own personally and that we read through your connected GitHub account, you can separately choose to let our server send Groq up to 8 selected, screened sentences from that repository's README, description or project documentation, plus technology names detected in its dependency files or reported by GitHub as its language, so Groq can suggest resume wording. This is a different consent from the public option above (version 3), has its own full disclosure, is never pre-ticked, applies to one import only, and is not available for organization-owned or shared repositories. Our screening also drops sentences that mention your GitHub account name, internal host names, or wording such as confidential or customer; this filtering is best effort, and a sent sentence can still contain names mentioned in your private README or documentation, such as the project's name. We do not send source code, whole files, file names, the repository link or full name, your GitHub account name, GitHub tokens, your resume or account details. Groq processes the information sent to it on its servers and may temporarily retain inference data under its own policy unless Zero Data Retention is enabled for our account. Before the suggestions are returned, we store coded fingerprints of them and the repository identity (not the raw text), so that AI features you use later can block that text unless you give separate permission for the repository. You review and choose any suggested lines before they become saved resume content.
We do not sell your data or use it for advertising.

Connected GitHub repositories

Connect GitHub is available only when enabled for your account. On GitHub, you may grant the MyResumy GitHub App access to selected repositories or to all repositories in an account. The App receives read-only access to contents and basic repository details for every repository covered by your choice. We list the repositories GitHub makes available through your connection. We store your GitHub account id and username and encrypted App authorization tokens in our database until you disconnect, delete your account, or we discover on a later connection check that GitHub has revoked or expired the authorization. Tokens are not sent to your browser.

Private repository import is available only when enabled for your account. GitHub receives the file-read requests, and our server runs on our hosting provider's infrastructure. An ordinary private import reads a selected README, project documentation and manifest files, plus file names and repository details. README, documentation and manifest files may contain code snippets or scripts. Optional private code analysis separately reads eligible source-code files only when it is enabled and you tick its separate box for that import. In that case, our server reads up to 6 selected source-code files and up to 3 manifest files, scans them for secrets, never runs them, and keeps the files and short descriptions extracted from them only in temporary memory for that request. Source code and those descriptions are not stored in our database. The generated draft may include words from the repository or, with separate code-analysis consent, from its code. During private import, no private source code, repository metadata, extracted descriptions, draft bullets or other private-derived facts are sent to Groq or another external AI provider, unless you separately tick the optional private-repository wording option described under GitHub import above. The draft is built on our server and returned to your browser for review.

A private draft is not saved as a project unless you add it. If you do, its project description and the bullets you keep become saved resume content; deleting or editing that project changes the saved content. A private project is hidden from your public portfolio until you switch it on, and we warn you before publishing a resume that shows it. Separately, before a private draft is returned, we store coded protective fingerprints of its text, not the raw source code or draft, together with the repository's name, id and owner category. These records help our AI features recognise private-derived text. They may persist even when an import fails after fingerprint recording and no draft reaches you, such as when the monthly import limit is reached. You can request Forget for a repository under the conditions below.

Before an AI feature such as cover letters, ATS analysis and tailoring, job matching, resume import or writing help sends text to Groq, our server checks for recognised private-derived content. If it finds any, it blocks the whole request unless you separately allow external AI processing for that repository. This permission is optional and separate from GitHub connection and code analysis. It covers only the content imported from that repository so far; newly imported content requires permission again. You can withdraw permission at any time. The check runs before each request to Groq, including later steps of an AI feature already running; text already sent cannot be recalled. Organization-owned repositories and repositories owned by another person's account cannot receive this permission under our current policy. If the check cannot run, the AI request is not sent and AI features may be temporarily unavailable. Matching has limits: substantially rewritten text, very short fragments, or text copied in from outside MyResumy may not be recognised and is then handled like the rest of your resume.

You can disconnect in the import dialog and remove the App's repository access on GitHub. Disconnecting asks GitHub to revoke the token and deletes our stored copy, stopping further repository reads. It does not remove projects or other resume content you already saved; you can edit or delete those yourself. Forget affects protective fingerprints as described under Data retention & deletion.

Public portfolios

Publishing a resume makes it visible to anyone who has its share link — this intentionally includes your contact details (email, phone) and photo, since the portfolio page's contact form and mailto links rely on them. There is no public directory or gallery where published resumes can be browsed or discovered by strangers; only someone with the exact link can view it. Unpublishing removes it immediately.

Data retention & deletion

Deleting a resume removes it from our database. To delete your entire account and all associated data, contact us via the feedback widget and we will process the deletion within 30 days. Fingerprints of text imported from private GitHub repositories (including those kept after an import that did not complete) are kept while that text may still be in your resumes: they are deleted with your account, or when you choose "Forget" for a repository. Forget deletes them only if our automatic check finds none of that repository's text in the resumes, published resumes and job applications stored for your account. The check has the recognition limits described above (for example, heavily rewritten text may not be found), and it is not available for accounts with a very large amount of stored content. If your stored content changes while we check, or the check cannot be completed, nothing is forgotten. We cannot check copies that exist only in your browser and were never saved; if such a copy is saved after you choose Forget, it is no longer recognised. Deleted data can remain in backups kept by our infrastructure providers until those backups are rotated out; we do not promise immediate removal from backups.

Third parties

We rely on Supabase (auth & database), Paddle (payments), Groq (AI), and our hosting provider. Each processes data only as needed to provide their service.

Your rights

You can access, correct, export, or delete your data at any time — most of it directly in the app, the rest by contacting us. EU/UK users have the rights provided by GDPR.