Privacy Policy

Everything below is data you give us to build your career memory; we do not buy data about you or track you across other sites.

• Account details: your name, email address, and password. The password is stored only as a salted hash, never in plain text.
• Career history facts you add through text chat, voice conversations, or career memory editing: roles, projects, skills, education, credentials, and preferences.
• Links you choose to make between a skill and the work that demonstrates it, picked from your own career facts. These are candidate-provided examples; we never present them as employer-verified.
• Voice intake audio and the conversation transcripts derived from it.
• Resume files (PDFs) you upload for fact extraction.
• Content the product derives from your facts: generated resumes, grounded application materials, job reviews (for each requirement a job lists, the verdict, the facts it rests on, and a text-similarity value), and application records. An application record can include status and timeline events, private notes and contacts, submitted material, and the source and time you confirm.
• Immutable application evidence snapshots: the exact stored job and resume versions, selected export check evidence, and bounded copies of the career-fact excerpts, generated artifacts, and contact involvement needed to show what you say was submitted. Private jobs you capture also retain their owner-scoped company identity and assignment history.
• Bounded export check reports for resume exports, including the format, an exact file fingerprint, recovered-field counts, and any issue codes, short labels, or bounded expected-content excerpts needed to explain a missing or reordered result. We do not store the exported file or its recovered full text.
• If you connect the Chrome extension: the account-bound connection record; the current job page URL and title; job title, company, description, location, work setup, employment type, source, canonical, and apply links; observation time and extraction method; your edits to those fields; and explicit save, generation, Application Assist, and submission-confirmation actions. The extension does not collect general browsing history, cookies, anything you type into a form except an answer you ask MemoryCV to remember, screenshots, or raw page HTML; if you explicitly choose “Read this page with AI”, or use a grounded Application Assist feature on a form whose job the companion cannot otherwise identify, the visible posting text (with form fields, hidden content, and page chrome removed) is sent once for extraction and is not retained.
• If you ask the agent to fill an application for you: a record of each hosted-browser session, and a per-turn history of the run covering why it stopped, the kinds of actions it took, how many fields it declined to fill, and the AI budget it used. These hold no field values and no screenshots.
• Answers you ask MemoryCV to remember, including any demographic, disability, veteran-status, or background-check answer, until you delete them.
• Answers you give in the Career Agent to an employer’s demographic, disability, veteran-status, or background-check question, stored with that application’s materials. They are kept for later applications only when you choose Save for next time.
• If you turn on alerts in Settings: the Web Push subscription your browser issues, which is a delivery endpoint, its two encryption keys, and your browser user agent. It is what lets us send an alert to that device.

Your data is used to run the product: extracting career facts, generating resumes that stay grounded in those facts, matching you against job postings, and tracking your applications. We do not sell personal data and we do not use it for advertising.

Fact extraction, resume generation, job matching, reading a job posting you link in the Career Agent (the page at your link is retrieved by Anthropic’s web fetch tool), and the Chrome extension’s page-read fallback are performed by external AI APIs (Anthropic for text processing, web fetch, and page-read extraction, OpenAI for embeddings, and Mistral for job requirement extraction and form field classification). Under those providers’ API terms, API inputs and outputs are not used to train their models by default. Their published standard API terms may retain inputs and outputs for up to 30 days for service and abuse-monitoring purposes, subject to our account settings, a different agreement, or legal requirements. Voice conversations are handled by ElevenLabs, which processes the audio and produces the transcripts we extract facts from.

When you ask the agent to fill an application for you, MemoryCV opens the employer’s site in a hosted browser run by Browserbase. Your selected resume file is staged into that browser for the upload field, and the answers you approved are typed into the form there. The employer’s site can receive the file and those answers as soon as they are entered, before you send the application. While it fills, no AI model sees a screenshot of the page or what is typed into it. To match your answers to the form, only the text of its fields goes to TypeSafe, or to Mistral when TypeSafe cannot answer. The agent fills a demographic, disability, veteran-status, or background-check question only with your own answer to that exact question, given in the Career Agent or saved earlier, never with a guess, and it never confirms, signs, or submits the application for you.

MemoryCV Companion installs with access to https pages so it can spot job postings and application forms wherever you browse, with no per-site setup. That detection runs entirely on your device: it looks for a posting in the page and keeps a short-lived note, per tab, of which job that tab was reading. The note stays in Chrome session storage, is never sent anywhere on its own, and is deleted when the tab closes or the browser exits. Nothing is sent to MemoryCV unless the side panel is open for that window or you take an explicit action, and nothing is saved to your account except by your explicit actions: generating a tailored resume, saving a job for later, using a grounded answer, or confirming a submission.

On application forms, the companion identifies form fields without reading what you have typed into them. An answer is stored only when you explicitly ask MemoryCV to remember it, limited to work authorization, sponsorship, compensation, relocation, availability, and the demographic, disability, veteran-status, and background-check questions you choose to save, and you can view and delete remembered answers from the panel at any time. The companion reads what you typed into one of those questions only when you ask it to remember that answer. It never confirms, signs, or submits an application for you.

Before you save, the extracted job fields are sent to MemoryCV so Mistral can extract requirements and OpenAI can embed those requirements for comparison with the connected career memory. This pre-save check does not create a saved job. The extension gets at most 24 requirement rows and evidence snippets capped at 140 characters, without career memory fact identifiers or raw facts.

Chrome local storage holds one scoped connection token and an isolated workspace for each open Chrome window. Depending on the current step, a workspace can contain page and job fields, observation metadata, bounded requirement and evidence rows, the saved job reference, current application stage, and resume-generation progress or destination path. Closing a window deletes its workspace. The token can save and score jobs, generate for the connected career memory’s own captured jobs, project career-memory-safe Application Assist details, export a selected resume for attachment, author grounded answers, save and manage the remembered answers described above, and confirm submission evidence. It cannot perform general application-stage writes, list or read career memory facts, resume text, application notes or history, billing, settings, or other account data.

MemoryCV Companion’s use of information received from browser permissions complies with the Chrome Web Store User Data Policy, including the Limited Use requirements. We use and transfer that data only to provide or improve the companion’s disclosed job capture, resume generation, application assistance, and tracking purpose. We do not sell it, use it for personalized advertising or credit decisions, or let people read it except with your specific consent for support, when needed for security or law, or as aggregated and anonymized data for permitted internal operations.

We use a small set of service providers. Each receives only what its column describes, only to provide that service.

We set a single essential session cookie so you stay signed in, and two preference cookies, each kept for a year, that remember how you left the sidebar and whether Applications shows as a board or a list. We also set one first-party analytics cookie, ph_<project>_posthog, holding an anonymous device identifier and PostHog’s own session state; it expires 365 days after your last visit. Those analytics requests are sent to a MemoryCV path on this domain and forwarded to PostHog. There are no advertising cookies. Company logos load in your browser straight from Brandfetch, as described above, or from the site that hosts that logo, which receives your IP address and browser user agent, never the page you are on; MemoryCV does not use these requests for analytics or advertising.

Your data is kept for as long as your account exists. Facts, resumes, and applications you delete inside the product are removed from your career memory. A confirmed submission is historical evidence: deleting or editing a source fact or resume does not silently rewrite the bounded copy already frozen in that application snapshot. Those copied excerpts remain until you delete the containing application or your account. Working resume export check reports are shorter-lived: we keep no more than the newest three for the same document version and analysis method, and delete all such working reports after 30 days. Facts extracted from an uploaded resume wait for your review: if you never accept that review, we delete the extracted review facts and the upload record after 30 days. A Chrome extension token expires after 90 days and can be revoked sooner in Settings. Disconnecting the extension clears its token and all panel workspaces from Chrome. Uninstalling clears the extension’s local data and permissions. Saved jobs, resumes, and application records remain in your MemoryCV account until you delete the relevant record or account. Providers listed above retain data under their own published policies. For voice intake that includes transcripts held by ElevenLabs.

You can delete your account yourself at any time from Settings. Deletion cancels active billing first, then removes your application evidence in an explicit order before removing your account and its remaining data: facts, voice transcripts, uploaded files, resumes, resume export check reports, owner-private job and company records, extension connections, job matches, applications, hosted-browser sessions and agent run history, registered alert devices, sessions, and any subscription.

You can download a machine-readable JSON copy of your data yourself at any time from Settings. For help with an export, or to ask what we hold about you, email support@memorycv.com from your account address.

If this policy changes, we update this page and the date at the top. Material changes will be flagged in the product before they take effect.

Privacy questions go to support@memorycv.com.