Privacy Policy
Last updated:
Information and purposes
CVFit processes account emails, display names, authentication and security records, and the resumes, versions, job descriptions, matches, applications, notes and follow-ups you provide. This supports editing, matching, job discovery and application management. Original uploaded resumes are stored encrypted for re-import; this does not mean all parsed text or workspace content uses the same encryption. We do not sell personal information.
Parsing and optional AI structure checks
Ordinary uploads, text extraction and structure parsing run within CVFit without automatically calling external AI structure mapping. Only selecting “Use AI to check structure” for a particular Studio re-import permits sending resume sections, stable block identifiers, source text, page order and limited font cues to the currently configured AI provider.
The structure check omits structured headers, contact fields and full layout coordinates, and excludes detected name and contact source lines. Companies, schools, qualifications and experience text remain necessary to validate mappings. Personal information can still appear in body text; this is not anonymization. Leave the option off to use local parsing, editing and PDF export.
AI reviews, Assistant and providers
When you request an AI resume review, detected personal information in the resume is automatically cleaned before relevant text, the job description and target role are sent to the selected provider. This does not mean job descriptions, free text or every AI feature is anonymized. Cleaning and generated content can be inaccurate. Avoid passwords, API keys, identity documents, financial or health information.
Using Assistant may send chat messages, necessary conversation history, page context and relevant resume, job or application evidence retrieved by tools to the provider. Conversations, sources, tool activity and approvals are stored within your account for continuity and operation review. Writes such as resume or application changes require confirmation. When a semantic retrieval model is configured, background tasks create or update indexes for ready resumes and saved versions even when chat is closed, sending relevant snippets to the configured embedding service; retrieval can also send query text. Without a configured model, keyword retrieval makes no embedding requests. Indexes are account scoped.
You may configure a compatible provider or, when eligible, use the managed service. Saved API keys are encrypted, are not shown again and are removed on provider reset or account deletion. Usage logs record models, usage and execution information rather than complete AI request bodies; conversations and workspace results are stored separately. Processing locations, retention and terms depend on the actual provider. We do not promise processing only in your country or that providers never retain content. Review the chosen provider’s terms before use.
Assistant memory and cross-conversation preferences
Conversation memory keeps goals, preferences and excerpts of completed requests within the current chat. You can turn it off or clear it in Assistant. Cross-conversation preferences are a separate setting, off by default. Only the reply language, response detail and career guidance preferences you explicitly save are stored in your account. Assistant may offer a choice based on a preference you explicitly state in the current message. Only choosing Remember or saving in the menu writes a long-term preference. Proposals and receipts remain in that chat; choosing Just this chat does not enable cross-conversation preferences or import other transcripts.
When enabled, these preferences are sent to your currently selected AI provider with each reply request. You can review, edit, pause or clear them. Pausing retains the saved content. Clearing removes preference content and resets the settings, while retaining revision state needed to prevent stale updates. Changes apply to subsequent reply requests, not a response already in progress. Existing messages and conversation memory are managed separately. Preferences are not evidence of resume experience and do not automatically change job search settings or approve actions.
Free text receives best-effort personal-information cleaning, not guaranteed anonymization or detection of every secret. Do not include passwords, API keys, contact details or other unnecessary sensitive information. Account deletion removes the associated preference record. Clearing or deletion cannot recall content already sent to providers or immediately rewrite backups; the deletion and retention provisions below also apply.
Assistant trial allowance
To track eight free messages per verified account, we store account-linked request identifiers, response status and timestamps, without chat text or API keys in the allowance records. A message counts once when answer text or a suggestion card starts being provided, regardless of tool calls. Stopping, disconnecting or a later failure does not restore it; a failure or stop before any such output releases the allowance. Deleting chats or refreshing does not reset the allowance. Account deletion removes its allowance records.
Public web, jobs and account security
Web searches and job discovery send necessary queries, locations, job preferences or URLs to the relevant search services and public sites. Results and sources may be saved. Job requirements are not evidence of your experience, and viewing jobs does not submit applications. Account emails use an email service; when challenges are enabled, Cloudflare Turnstile processes browser and connection information needed for verification. Security events, hashed network/browser identifiers and short-lived quota state help prevent abuse; hashing is not complete anonymization. Authentication cookies and locally saved language preferences support sign-in and interface functions.
In-app feedback and support email
When feedback is available, signed-in submissions record category, title, description, account association, timestamps and hashed abuse-prevention identifiers. The account email at submission is saved for that ticket only if you allow contact. Authorized administrators may review, handle and add internal notes; status changes do not automatically send emails. A receipt is not a public ticket-reading capability.
Administrators may manually export feedback through a field whitelist and text cleaner for issue analysis or controlled AI assistance. Contact email and internal notes are excluded by default and require separate selections. Account IDs, network and session security identifiers are excluded from exports. Free text may still identify people: cleaning is not complete anonymization, and administrators must review before sharing. Export does not automatically transmit data to external AI.
Support email sends your sender address, subject and message through the actual mailbox provider to handle your request. Email is not automatically imported into the ticket database. Do not include passwords, verification codes or unnecessary sensitive information.
Deletion, retention and access
Account settings let you delete your account and associated workspace data. Support feedback still linked to that account is also deleted, including contact emails and internal notes. Feedback historically unlinked from an account cannot be reliably attributed automatically; contact support with a receipt or other necessary details so we can assess your request.
Database deletion cannot recall downloaded exports, sent emails or copies already held by third parties, and does not immediately rewrite every backup. Those copies require separate handling. We do not promise a uniform 180-day or 30-day period; information is retained only as needed for the service, security, request handling or applicable law. Contact us to request access, correction or deletion, or to ask about processing. Identity verification may be necessary.
Results and contact
Match scores, ATS indicators and AI suggestions support your judgment; they are not hiring decisions or employment guarantees. For privacy, support or deletion requests, use Login help or email support@cvfit.online. Material policy changes will update this page’s date and explanation.
Tracker exports and calendar subscriptions
Downloading records or a single calendar event is initiated by you; notes are excluded by default. Enabling a follow-up calendar subscription lets your calendar app or its provider use a private link to read the company, role and date for every record with a follow-up date. Resumes, job descriptions and notes are excluded. Anyone with this link can read these calendar details. Do not publish or forward it.
Subscriptions do not read your existing calendar. Changes in CVFit appear when the calendar service refreshes and may be delayed; calendar edits do not update CVFit. You can disable the link, and deleting your account invalidates it. Disabling cannot recall copies already saved by a provider or device; unsubscribe in your calendar app as well. Static file imports do not receive subsequent changes.