Privacy Policy

Effective 2026-11-07. This is the formal Privacy Policy for Krewva's iOS, Android, macOS, and Windows applications. The formal Privacy Policy controls if this page differs from the Trust Charter. It explains what Krewva sees, what stays on the device, what may reach Krewva and the provider for the feature you choose, and how long records are kept.

Mobile Reply provider notice updated 2026-11-07. Reply requires its separate 2026-10-22 acknowledgement. Signing in accepts the Terms and Privacy Policy only. A separate in-app AI choice describes the third-party LLM processing and records the permissions you explicitly accept. The general Terms and Privacy acceptance remains version 2026-09-04. Voluntary support feedback and iOS Dictate cleanup retain their separate notices. Their notices appear before you submit feedback or start Dictate; Dictate records its updated Voice acknowledgement.

Who Krewva is

Krewva is made by Krewva Inc. (formerly Wuvov Inc.), a Delaware corporation. “Krewva” means the product; “you” means the person whose accounts, device, or content Krewva uses under your instructions.

What Krewva collects

Krewva collects the smallest set of information needed for an enabled feature and the records needed to show its work. Depending on what you use, this can include:

  • Account basics, consent and feature settings, permissions, approvals, and request or audit identifiers.
  • Content from the connected accounts and local context that you authorize, such as messages, threads, contacts, and selected files.
  • Derived records such as drafts, review evidence, Task candidates, accepted Task specifications, plans, outcomes, and selected Task evidence.
  • Content-free diagnostics such as platform, readiness, request ID, dimensions, byte size, timing, failure code, and candidate count.

Krewva does not collect connected-account content for advertising or track you across the web. A screenshot can contain personal data about you or other people, so feature-specific minimization does not mean that every identifier will be removed.

Mobile product analytics

Core app and keyboard-quality analytics are enabled by default during this beta after legal onboarding on iPhone and Android, when collection is available and the account is eligible. AI processing, keyboard Full Access and microphone permissions remain separate. Deletion requests remain protected across sign-in.

Core analytics records fixed page and button identifiers, foreground time, feature attempts, explicit reply choices, insertion calls, fixed failure codes, and app and operating-system versions. Available events depend on the feature and platform. Earlier clicks are not reconstructed. Events use a random account-linked identifier, not an anonymous identity. This beta does not request age, occupation or relationship research questionnaires. Krewva does not scan installed apps or infer demographic answers from conversations or contacts.

This analytics path excludes conversation and reply text, prompts, screenshots, audio, keystrokes, clipboard contents, contact identifiers, names, email addresses, advertising identifiers, and device fingerprints. It uses no session replay or automatic location enrichment and does not forward your client IP to the analytics provider. AI requests remain a separate path and may process the screenshot described below.

The selected analytics provider is PostHog Cloud, through its US endpoint. During this beta, we do not set an event expiry date; vendor retention limits may apply. Inactive research profiles and analytics identities are scheduled for deletion after 12 months without permitted activity. Product and engineering maintainers may inspect these records. Keyboard activity is used only to improve the keyboard. Marketing receives only reviewed non-keyboard summaries, with small groups withheld; it receives no platform account or individual advertising audience. This does not make the underlying records anonymous.

Contact privacy@krewva.com to request analytics deletion. Deletion stops collection and requests removal of the provider profile and events, separately from records the law requires us to keep. Requests already in flight may finish. We verify completion, with a target of 30 days.

Mobile Reply and desktop screen flows

Screen access, feature enablement, AI-processing consent, Task acceptance, Step approval, and outbound action authority are separate. Krewva does not treat an operating-system screen permission as feature enablement or AI consent. Task acceptance and Step approval do not silently grant screen access or pixel upload authority.

  • Mobile Reply. Reply on iOS and Android uses a third-party LLM provider. Before a Reply request is sent, Krewva asks for separate Reply permission. On iOS, a separate in-app AI choice after sign-in describes the third-party processing; agreeing to the Terms and Privacy Policy alone does not grant AI permission. Android retains its feature-specific choice. The request may include the current conversation text or screenshot, your reply instructions, and selected personalization context, such as relevant profile, contact and writing information. You review the drafts and choose what to insert; the iPhone app offers five drafts. Krewva never sends a message for you. Basic typing does not itself start a Reply request.
  • Desktop Draft a Reply. A user trigger requests one fresh screenshot of the active top-level window. Only when the feature is enabled and AI processing is consented to may that screenshot travel through Krewva to a third-party LLM provider for reviewable reply suggestions. The capture itself never sends.
  • Desktop Do a Task local OCR. Krewva tries semantic text first. If a bounded semantic gap remains, bounded local OCR can process active-window pixels. Raw pixels stay on the device. Selected derived text may support a Task after the user's choice, not before.
  • Desktop Do a Task screenshot discovery. This is a separate enabled and AI-consented upload path. A fresh active-window screenshot may travel through Krewva to a third-party LLM provider to propose Task candidates only. A candidate grants no execution authority. Task acceptance and Step approval remain separate.

On macOS and Windows, capture is user-triggered and limited to the active top-level window. It is not continuous monitoring and never falls back to a full-desktop capture. Denied, revoked, protected, or changed targets fail closed. Windows may request borderless-capture access during explicit setup or Settings; a normal trigger does not open that prompt. The platform permission mechanisms differ while the product meaning stays the same. You are responsible for lawful access and any notice or consent required for other people visible in a window.

Providers and their data rules

Voice Input, Dictate, Intent, Refine and independent personalization learning use third-party LLM providers under their existing permissions. Desktop AI processing is unchanged. Agreeing to the general AI notice alone does not grant Reply permission. The separate iOS in-app AI choice describes the third-party processing and records both permissions only when you explicitly accept. You can review or revoke these under Settings → Account → Privacy. Turning off AI drafts revokes the general AI and Reply permissions; Voice Input has its separate permission.

On current iPhone and Android versions, ordinary Dictate and spoken Intent send microphone audio through Krewva to a third-party speech and language provider while you record, after the required voice-processing consent. Stop finishes the request. Cancel stops recording; data already transmitted remains subject to provider retention. Spoken Intent can include relevant saved writing guidance and personal context. On Android, permitted conversation access can also supply relevant conversation text. Krewva does not save raw recordings in History; saved text follows the history and backup rules below. Where saved voice-processing permission is offered, it applies to your account and the current notice. Revoking permission or changing the notice requires fresh agreement.

Conversation screenshots and text can include information about other people. Selected personalization context can include names and relevant saved information. Krewva does not promise that every identifying detail is removed before this processing.

Krewva's third-party LLM providers process requests under their own terms. Under the API terms Krewva relies on, prompts and responses are not used to train or improve the providers' models unless Krewva explicitly opts in. Providers still retain request data for abuse monitoring for a limited period (currently described as up to 30 days), may review flagged content, and may process data in countries where they or their agents maintain facilities. Krewva does not promise zero provider retention or processing only in the United States.

For website hosting, waitlist processing, KV storage, and rate limiting, Krewva uses Cloudflare. Other operational and connected-account providers are listed in the sub-processors list.

Screenshots, diagnostics, and permitted derivatives

Screenshots captured for AI features are request-scoped. Krewva excludes them from account history, ordinary logs, Sentry, and support bundles. Krewva may retain permitted derivatives under the relevant feature rules, including generated drafts, review evidence, selected Task evidence, plans, outcomes, audit facts, and a one-way fingerprint. Those derivatives are not raw screenshot history.

When you submit feedback, your message and screenshots you explicitly select are sent through Krewva to a private support channel in Slack, together with your Krewva account ID, name, email address and app version. These support attachments remain in Slack under the support workspace's access and retention settings; they are not automatically removed by Krewva account deletion. Feedback does not include automatic screenshots, recordings, conversation history, or diagnostic logs.

Diagnostics are designed to remain content-free. They may record platform, readiness, request ID, image dimensions, byte size, whether a fingerprint exists, timing, failure code, or candidate count, but not raw screenshot pixels.

How long Krewva keeps information

Desktop message bodies, cards, and voice samples are purged after 365 days of account inactivity; Krewva has one window for every account. It does not apply to AI-feature screenshots, which are request-scoped and excluded from Krewva account history. Consent-gated redacted improvement records and audit records have separate lifecycles and are not raw message history.

Voice Input is separate from AI drafts. The app asks before microphone audio is sent to a third-party speech-to-text provider for transcription, and Krewva does not treat AI draft consent as Voice Input authority. Final recognized text may pass through Krewva and a third-party LLM provider once for faithful cleanup of fillers, repetition, and explicit self-corrections. This cleanup request does not itself create server-side transcript history; saved history may be backed up separately as described below.

In the iPhone or Android app, Settings → Account deletion starts deletion immediately and signs you out after the request is accepted. No 30-day grace period applies to this in-app path. Purging is automatic; a completion notice follows when the deletion finishes. Older desktop clients may still schedule deletion with a cancelable 30-day grace period. Third-party access revocation is best effort. Legally required or separately governed records may remain for their stated lifecycle. Krewva does not promise that a provider or connected account will delete data on the same schedule.

You can also request account deletion without opening the app. Email privacy@krewva.com from your registered address with the subject “Delete my Krewva account,” including saved history and personalization backups in your request. Krewva verifies account ownership before processing. Do not send passwords or verification codes.

iPhone Reply History, Writing Samples, Contacts, Objectives and Personal Context are stored on your device. When you are signed in and the app can connect, saved history and personalization can also be backed up automatically to your Krewva account. This includes saved Voice and Intent text, relevant conversation text, learned preferences and observed final-edit text fragments. Account backup is separate from the selected context sent to an AI provider for a request. Reply History, Writing Samples and Contacts remain until you delete them. Selected context may be sent through Krewva to a third-party LLM provider for Reply or for the other AI features described here, after the relevant consent. Account deletion clears this local personalization data. The stored Writing Samples archive is separate from the bounded set selected for each request.

On Android, History, Writing Samples, Contacts and Personal Context are saved on your device and can be backed up automatically to your Krewva account when signed in and connected. Saved records can include Voice and Intent text, conversation text, drafts, learned preferences and observed final edits. History remains until you remove it. Account backup is separate from the context selected for an AI request.

Android writing learning starts on for new local profiles. Original text typed with Krewva or dictated may become Writing Samples after a pause or completion; unchanged AI replies and Intent output are excluded from original-writing samples. Saved samples can leave the device through backup even when basic typing does not start an AI Reply request. Selected saved context may also go to an AI provider after the relevant consent.

On Android, you can delete saved History entries and Writing Samples, edit or delete Personal Context, forget saved contact items, or clear all personalization. Contact learning has separate pause and resume controls; forgetting an item does not itself pause future learning. Backup changes finish when connected. These controls are separate from deleting your account, which also clears local personalization.

How Krewva protects information

Krewva authenticates account requests and encrypts connected-account credentials at rest with keys managed by AWS KMS. Access is bounded by the feature, consent, Task, Step, and destination authority that apply to the request. When a required permission or consent is denied, revoked, or stale, the relevant operation fails closed.

Your choices and requests

You can disable a feature, revoke an operating-system or connected account permission, reject a Task candidate, decline Task acceptance, or decline a Step approval. You can request access, correction, deletion, or an export of the information Krewva keeps, subject to identity checks, product availability, and any applicable legal or separately governed record requirement. Contact Krewva if the in-app path does not cover your request.

Children and other people's content

Krewva is not designed for children to use independently. Do not connect or capture another person's content unless you have lawful authority and provide any notice or obtain any consent that is required.

Changes to this notice

If this notice changes materially, Krewva will update the version date and use its normal acceptance gate before the revised policy applies to existing users where required.

How to reach Krewva

Questions about privacy or a request about your information — write to privacy@krewva.com. A person reads it.