Privacy Policy

This Privacy Policy explains how Prelto collects, uses, and protects information when you use the Prelto Mac application, websites, and cloud services.

1. Information We Collect

Depending on the features you use, Prelto collects account information (email address, display name), authentication data, billing status, device metadata (a randomly generated installation ID, device name, macOS version, and app version), content you choose to sync (snippets, custom vocabulary, and writing instructions), metadata about AI requests, and support messages you send us. The Mac app also processes content locally on your device, including microphone audio, transcripts, on-screen context, and generated text.

2. Voice, Screen, and Text Content

Speech-to-text runs entirely on your Mac using a local model. Raw microphone audio is never sent to Prelto servers. The app briefly stores audio on your device for crash recovery only; that buffer is deleted when a capture completes and is automatically removed within 24 hours in all cases.

When you dictate with cloud formatting enabled, the transcript text (and your custom vocabulary, if enabled) is sent to Prelto Cloud for polishing. When you ask a question or request a draft, Prelto sends what it captured at the moment you pressed the key: your spoken request, the frontmost app name, window title, browser page address, selected text, focused-field text, and, when needed to answer, a screenshot of the active window (or the active display if a window capture is not available). Screenshots are skipped when your selection is sufficient or when Screen Recording permission is not granted. Context is captured only at the moment you press the key; Prelto has no ambient or background capture.

Prelto's production database does not store transcripts, prompts, selected text, screenshots, or model outputs. Cloud AI usage records are metadata only: route, model, character and token counts, status, timing, and cost accounting.

The app keeps a local history of recent activity on your Mac (up to 100 entries for up to 30 days), which can include transcripts and selected text. You can turn retained content off in settings, which also removes previously retained text. Diagnostic logs on your device redact content by default and are never uploaded automatically.

A hidden, off-by-default diagnostics option lets you upload the full content of a specific request (including its screenshot) to our debugging tools. It only applies when you explicitly enable it.

3. Accounts, Billing, and Entitlements

Prelto uses account and billing information to authenticate users, manage access, apply trials or beta grants, enforce usage limits, provide subscriptions, and prevent abuse. Payments are processed by Stripe; your card details go directly to Stripe and never touch Prelto servers.

4. Sign in with Google or Apple

If you choose Sign in with Google, Google provides Prelto's authentication service with your Google account identifier, email address, email verification status, and basic profile name. Prelto uses this information only to create, authenticate, or link your Prelto account. Prelto does not request access to your Gmail, Google Drive, Google Calendar, contacts, or other Google content through this sign-in flow, and does not sell Google user data. Sign in with Apple works the same way with your Apple ID email and name.

5. Sync

If sync is enabled for your account, your snippets, custom vocabulary, and writing instructions are stored on Prelto servers so they follow your account. Turning sync off deletes your synced snippets and vocabulary from Prelto servers.

6. Service Providers

We use a small set of service providers: Amazon Web Services (authentication via Amazon Cognito and delivery of verification emails), Stripe (billing), OpenRouter (AI processing, configured to deny provider data collection and to require zero-data-retention endpoints), Braintrust (service monitoring that receives request metadata such as model, latency, and token counts; request content only if you enable the diagnostics option above), Google Cloud (API hosting and database), and Cloudflare (website hosting, downloads, and update delivery). Automatic update checks contact our download server and expose only your IP address and app version, with no system profiling.

7. How We Use Information

• Provide, secure, and improve Prelto.
• Transcribe, transform, and format user-requested content.
• Operate account, billing, trial, beta, and entitlement systems.
• Respond to support requests and diagnose product issues.
• Comply with legal, security, and abuse-prevention obligations.

Prelto does not sell your data and does not use your content to train AI models. The Mac app and website contain no third-party analytics or advertising trackers.

8. Data Retention

We keep account records while your account is active and as needed for legal, security, billing, backup, or operational reasons. Cloud AI usage records are metadata only, as described above. Local data stored by the Mac app stays on your device unless you enable sync or the diagnostics option.

9. Security

We use technical and organizational safeguards designed to protect information in transit and at rest. No system can be guaranteed perfectly secure, and you are responsible for maintaining the security of your devices and account credentials.

10. Your Choices

You can use local formatting instead of cloud formatting, turn off retained history on your Mac, disable sync (which deletes synced content from our servers), cancel your subscription at any time, delete local app data from your device, and contact us to request deletion of your account data.

11. Children

Prelto is not directed to children under 13, and we do not knowingly collect personal information from children under 13.

12. Changes

We may update this policy from time to time. The latest version will be posted on this page with an updated date.

13. Contact

Questions or requests can be sent to [email protected].