Privacy Policy

VyraScribe ("we," "our," or "us") is operated by Kyle Kendrick, d/b/a VyraScribe. This Privacy Policy explains how we collect, use, and protect your information.

What We Collect

• Account information: email address, name
• Profile data: the 10 profile fields you provide during onboarding (role, audience, offer, industry, voice, goal, platforms, experience level, custom instructions)
• Voiceprint data: when you build your Voiceprint, your writing samples and voice recordings (as text transcripts) are stored in your account so you can view, manage, and add to them at any time. You can delete individual samples or your entire Voiceprint from the My Voice page. Deleted samples are permanently removed immediately. Your samples are processed by Anthropic's Claude API during Voiceprint analysis (disclosed under our existing Anthropic subprocessor disclosure) and stored in our Cloudflare D1 database (disclosed under our existing Cloudflare subprocessor disclosure). Raw audio recordings are transcribed and immediately discarded — only the text transcript is stored.
• Generated content: content you generate and choose to save
• LinkedIn Studio inputs: when you use LinkedIn Studio tools (Idea Engine, Post Builder, VyraCheck), we collect the inputs you provide — topics, weekly context, tone and length preferences, and source material (stats, quotes, stories). For VyraCheck, your draft text and analysis results (verdict, scores, suggestions) are stored in your account so you can revisit and track improvements over time. You can delete any saved analysis from the VyraCheck page. All analysis data is processed by Anthropic's Claude API and stored in our Cloudflare D1 database.
• Feedback and quality data: when you rate a generation ("Sounds like me" / "Not quite") or provide written feedback, that rating and text are stored alongside the generation record. This helps us improve voice-matching quality. Feedback is voluntary and can be submitted once per generation.
• Analytics events: we record lightweight usage events (e.g., "copied post to clipboard") tied to your account. These contain no personal content — only the action name and a reference to the generation.
• LinkedIn performance data: When you enable performance tracking in the Chrome Extension, we collect public engagement metrics from your LinkedIn posts — including reaction counts, comment counts, repost counts, and reaction type breakdowns. This data is retrieved from your public LinkedIn profile using the Apify scraping service. Only publicly visible data is collected. Post content and metrics are stored in your account so Elvyra can analyze your content performance patterns over time. You can disable performance tracking at any time from the extension popup. Disabling tracking stops new data collection; existing data remains in your account until you delete it or delete your account.
• Elvyra chat conversations: When you chat with Elvyra on the VyraScribe web app, your messages and Elvyra's responses are stored in your account. Conversations are retained for 90 days and then automatically deleted. You can delete any conversation at any time from the chat panel. Chat messages are processed by Anthropic's Claude API for response generation (covered under our existing Anthropic subprocessor disclosure). Elvyra uses your profile, Voiceprint, and usage history to personalize responses — this data is assembled per-request and not stored separately.
• Usage data: generation counts, feature usage (aggregated, not individually tracked)
• Payment information: handled entirely by Stripe — we never see or store your card details

How We Use Your Data

• To personalize generated content based on your profile and Voiceprint
• To analyze writing samples and voice recordings during Voice Capture to build your Voiceprint
• To authenticate you and maintain your session
• To process payments and manage subscriptions (via Stripe)
• To send transactional emails (magic links, account changes) via Resend
• To generate LinkedIn content (posts, ideas, audits) personalized to your Voiceprint and profile via LinkedIn Studio
• To search the web on your behalf for current industry topics when using Idea Engine's research mode (via Anthropic's web search capability, covered under our existing Anthropic subprocessor disclosure)
• To improve voice-matching accuracy using your voluntary feedback ratings
• To power Elvyra's conversational chat, which provides personalized writing insights, suggestions, and assistance based on your profile and Voiceprint via Anthropic's Claude API
• To improve the service through aggregate usage analytics (no individual tracking cookies)

Third-Party Services (Subprocessors)

• Cloudflare: Infrastructure, hosting, CDN, D1 database. Privacy Policy
• Anthropic: Content generation engine. Your profile data and prompt content are sent to Anthropic for each generation. Anthropic does not train on data sent through the API. Privacy Policy
• Resend: Transactional email delivery. Receives your email address. Privacy Policy
• Stripe: Payment processing. Handles billing address and payment method. PCI-DSS Level 1 compliant. Privacy Policy
• Apify: LinkedIn post scraping for performance tracking. Receives your LinkedIn profile URL and returns publicly visible post data (post content, engagement metrics). Apify does not receive your VyraScribe account information. Privacy Policy

Chrome Extension

When using the VyraScribe Chrome Extension, your assembled prompt (which includes your profile and Voiceprint data) is inserted into the chat tool you're using (such as ChatGPT or Claude). VyraScribe does not control how that tool processes your data. Review the privacy policy of the tool you're using for details on their data handling.

When you enable LinkedIn performance tracking, the extension detects when you visit your own LinkedIn profile and triggers a background data sync. This sends your LinkedIn username to our server, which uses the Apify service to retrieve your publicly visible post engagement data. This data is stored in your VyraScribe account and used by Elvyra to provide content performance insights. No LinkedIn login credentials are accessed or stored.

Cookies

We use only strictly necessary cookies: a session cookie for authentication. We do not use tracking cookies, advertising cookies, or analytics cookies. Stripe.js may set essential cookies during checkout, classified as necessary for payment processing.

Data Retention

• Saved outputs: kept until you delete them
• Unsaved generations (including LinkedIn Studio outputs): retained for 30 days, then automatically purged
• LinkedIn performance data: retained for the lifetime of your account or until you disable performance tracking and request deletion. Metric snapshots are collected at up to 6 intervals per post (from 1 hour to 14 days after publication).
• Elvyra chat conversations: retained for 90 days, then automatically deleted. You can delete individual conversations at any time.
• Feedback ratings and analytics events: retained for the lifetime of your account
• Account data: kept until you delete your account (30-day recovery window, then permanently deleted)
• Invoices and subscription records: retained indefinitely for tax compliance

Your Rights

• Access: view your data in Settings → My Profile and My Library
• Export: download all your data as JSON from Settings → Data → Export
• Deletion: delete your account from Settings → Danger Zone → Delete Account. After 30 days, all data is permanently removed.
• Correction: update your profile anytime via the Profile Builder

To exercise any right, email support@vyralab.com.

Security

We use HTTPS everywhere, encrypt data at rest (via Cloudflare D1), hash session tokens with SHA-256, and never store passwords (we use passwordless magic link authentication). All API keys and secrets are stored in Cloudflare's encrypted secrets manager.

Changes

We may update this policy. Material changes will be communicated via email or an in-app notice. Continued use after changes constitutes acceptance.

Contact

Questions? Email support@vyralab.com.