PRIVACY
Your voice, content, and data deserve protection.
Crafted Virtue helps professionals build public authority, which means we handle sensitive information: professional context, writing samples, generated drafts, connected channels, performance data, and brand preferences. This page explains how we think about that responsibility.
Last updated: September 2026
This is a general product policy page intended to communicate Crafted Virtue's approach to privacy. It should be reviewed by qualified legal counsel before production use.
1.What We Collect
We collect the information needed to create and operate your personal-branding workspace.
- Account information such as name, email, title, company, and login details.
- Professional profile information such as industry, goals, audience, biography, and public links.
- Voice and content inputs such as writing samples, approved posts, edits, comments, and draft preferences.
- Generated content such as drafts, revisions, media concepts, approval history, and publishing plans.
- Connected-channel metadata such as connected platforms, publishing status, and analytics returned from those platforms.
- Usage and product data such as onboarding progress, feature usage, support interactions, and app activity.
- Billing information handled through payment providers; Crafted Virtue should not store full payment card details directly.
2.How We Use Information
We use your information to operate the product and improve the quality of your personal brand system.
- Build and maintain your Voice Profile.
- Identify style preferences from your own edits to drafts, where your plan includes it.
- Generate, revise, and improve content drafts.
- Create publishing plans and approval workflows.
- Show analytics, Brand Score, Influence Delta, and performance briefings.
- Support onboarding, training, billing, and customer service.
- Detect errors, abuse, or workflow failures.
- Improve product quality and user experience.
2a.Research (beta cohort only)
If you are a founding-beta participant and you separately opt in, we also process a defined research extract to measure whether the service actually does what it claims — how your published voice, publication rate and editing burden change over twelve months.
The extracted fields, exhaustively: (1) your writing samples and their derived style measurements (such as sentence length and vocabulary frequency vectors); (2) counts and timings of drafts, edits, approvals and publications; (3) edit distance between drafts and your approved versions; (4) your answers to the enrolment and follow-up questionnaires, including the Willingness-to-Self-Censor scale; (5) wave timestamps. Nothing else enters the research extract. In particular: no message contents beyond your own drafts and samples, no third-party data about you, no platform analytics unless you separately supply them.
The extract is wave-stamped and stored separately. It is pseudonymised, not anonymised — we hold the key — and it is treated as personal data throughout. Published results are aggregated; no individual is identifiable in anything we publish.
Research participation is voluntary, separate from beta access, and withdrawable at any time. Withdrawal stops future extraction immediately; already-published aggregates are not retracted. One consent covers the full twelve-month programme — we will not re-ask at each wave.
3.Voice Profile and Writing Samples
Your writing samples and edits are used to help Crafted Virtue understand your style. The goal is not to replace your voice. The goal is to help your agent draft content that sounds more like you, while keeping you in control.
When you revise a draft written for you, we keep an exact copy of the text before and after your revision as evidence for your Voice Profile, together with when you made it, which draft it belongs to and a measure of how much changed. We also record, with the draft's text, when a draft written for you is approved, returned for changes or rejected. A draft you approve without editing is never used to identify a preference.
Where your plan includes preferences from your edits, Olivia, the specialist agent that builds your Voice Profile, reviews your substantial revisions with an AI model that we operate ourselves (see Section 6) and identifies changes you make repeatedly.
A change becomes a preference only after you have made it in several drafts on different topics, and new preferences are added gradually. Preferences in use are then applied automatically when new drafts are written for you. They cover wording and structure only: a preference is never used to change a factual claim, and it never approves anything.
Only edits you make yourself are used. Edits made by colleagues are not used, and your preferences are used only for drafts written for you, never for another person or organisation.
Your Voice Profile page shows each preference with the edits behind it. You can turn off any preference at any time, and allow it again later. A preference you allow again is not used until it has passed the same checks as a new one.
We use your revisions in this way to provide the service, as described in Section 2. Section 8 explains how long they are kept, and Section 9 lists your choices.
Nothing publishes until you approve it.
4.Connected Channels
When you connect publishing channels, Crafted Virtue may receive platform metadata and analytics required to schedule, track, and report on content. Publishing permissions can be revoked through your connected channel settings.
5.AI Processing
Crafted Virtue uses AI systems to help with onboarding, content drafting, voice learning, analytics, research, media planning, and support. AI outputs can be useful but should be reviewed. For user-facing content, approval remains part of the product experience.
6.Data Sharing
We do not sell your personal content or writing samples. We may share limited information with service providers that help us operate the product, including hosting, email delivery, analytics, billing, publishing, support, and AI-processing providers.
Preferences from your edits (Section 3) are identified by an AI model that we operate on servers we control. Our hosting provider supplies those servers; no outside AI provider receives your edits for this purpose.
7.Enterprise Workspaces
Enterprise customers may have administrators, brand managers, or reviewers who can access workspace activity according to their assigned permissions. Enterprise accounts may also require audit logs, content archives, approval records, and compliance exports.
In an enterprise workspace, preferences are identified only from each person's own edits to drafts written for them. Edits made by administrators, brand managers, reviewers or other colleagues are not used to identify anyone's preferences.
8.Data Retention
We retain account, content, analytics, and audit information for as long as needed to provide the service, meet operational needs, support enterprise compliance, or satisfy legal obligations. Where content was created for an organisation's account, the approval and publication evidence attached to that content — including who approved which exact version, and when — belongs to the organisation's record and survives the deletion of an individual member's account. If you are a founding-beta research participant who opted in under Section 2a, the research extract is retained for the study term plus 24 months for reproducibility, then destroyed or irreversibly aggregated.
The copies of your revisions kept as Voice Profile evidence (Section 3), and the preferences identified from them, are kept as described above. They are included when you download your data, and they are deleted when your account is deleted.
9.Your Choices
- Update profile information.
- Add or remove writing samples.
- Turn off, or allow again, any preference identified from your edits.
- Disconnect publishing channels.
- Manage notification preferences.
- Request data access or deletion where applicable.
- Cancel or modify your subscription.
10.Contact
For privacy questions, contact the Crafted Virtue team through the support channel or contact form.