Privacy Policy

1. What data we collect

When you connect your Google account, we request the following API permissions (scopes):

• gmail.readonly — read your inbox to identify emails needing replies
• gmail.compose — create draft replies in your Gmail
• gmail.modify — mark drafts as sent and manage labels
• calendar.readonly — read your calendar for scheduling context
• drive.readonly — read your Google Drive documents for knowledge context

Here is what we store:

• Email metadata — sender, recipient, subject line, timestamps, and thread identifiers for emails that require a reply.
• Email content — the body text of inbound emails, which we process through our AI pipeline to generate drafts. We do not permanently store raw inbound email bodies (see section 2).
• Sent email samples — a selection of your sent emails used during onboarding to build your voice understanding. These are analyzed and then discarded; only the resulting voice model is retained.
• Voice understanding — a structured representation of your writing style, tone, vocabulary patterns, and communication preferences derived from your sent emails.
• Contact voice profiles — per-sender models that capture how you typically respond to specific contacts, including formality level, typical response length, and relationship context.
• Claude conversation exports — if you choose to import conversations from Claude, we store those exports to incorporate additional context into your voice understanding.
• Drafts and learnings — the email drafts we generate for you, along with learnings we extract when you edit, send, or delete a draft. These learnings improve future drafts.
• Account information — your name, email address, profile photo (from Google), timezone, and subscription status.
• Calendar data — event titles, times, and attendee lists from your Google Calendar, used to provide scheduling context when generating drafts.
• Google Drive documents — document titles and content from files you've authored, processed to build knowledge context. Only documents you've created or edited are accessed.
• Payment information — handled entirely by Stripe. We store your Stripe customer ID and subscription status but never see or store your card number, CVV, or billing details.

2. What we don't store

We are deliberate about minimizing what we keep. The following data is processed transiently and never written to our database:

• Raw inbound email bodies — we pass these through our AI pipeline to generate drafts and extract context, but we store only the AI-generated summary and draft, not the original email text.
• Email attachments — we do not download, process, or store any email attachments.
• Emails you don't need to reply to — newsletters, promotions, and automated messages are filtered out and never processed.
• Browsing or device data — we do not use tracking pixels, fingerprinting, or analytics cookies. We do not sell data to advertisers.

3. How we use your data

Every piece of data we collect serves one purpose: writing email drafts that sound like you. Specifically:

• Voice analysis — your sent emails are analyzed by Anthropic's Claude API to build a voice understanding that captures your writing style. This analysis runs once during onboarding and is incrementally enhanced over time.
• Draft generation — when an email needs a reply, we send the email context, your voice understanding, and relevant learnings to Claude to generate a draft in your voice.
• Learning extraction — when you edit a draft before sending, or delete a draft, we analyze the difference to extract a learning. These learnings are stored and used to improve future drafts.
• Reply profile building — we analyze your historical reply patterns to understand how you typically respond to different contacts and email types. This helps us match your natural behavior.
• Scheduling — we use your timezone and preferred draft delivery time to determine when to generate drafts for you.

We do not use your data to train AI models. Your emails are processed via Anthropic's API, which has a zero-retention policy for API inputs and outputs. Your data is not used to improve Claude or any other model.

4. Third-party services

We rely on a small number of trusted infrastructure providers. Each processes only the minimum data necessary for their function:

• Google (OAuth & Gmail API) — authenticates your identity and provides read access to your Gmail inbox. Google's privacy policy governs their handling of your data. We request the minimum scopes necessary and use offline access tokens that we store encrypted in our database.
• Anthropic (Claude API) — processes email content and generates drafts. Anthropic's API has a zero-data-retention policy: inputs and outputs are not stored or used for model training. We use Claude Sonnet for user-facing tasks and Claude Haiku for background processing.
• Supabase — hosts our PostgreSQL database and provides authentication infrastructure. All data is encrypted at rest (AES-256) and in transit (TLS 1.2+). Our database is hosted in the United States.
• Stripe — processes subscription payments. Stripe is PCI DSS Level 1 certified. We never handle or store raw payment credentials; Stripe manages this entirely.
• Vercel — hosts our application and runs scheduled background jobs (cron). Vercel processes HTTP requests and logs but does not have access to your email content or personal data beyond what appears in standard server logs.
• Resend — sends transactional emails (welcome, trial notifications, weekly digest). Receives only your email address and email content we generate. Resend does not have access to your Gmail data.
• Voyage AI — generates text embeddings for semantic search across your knowledge base and voice understanding. Receives text snippets for embedding only. Voyage AI does not store inputs after processing.

We do not sell, rent, or share your personal data with any other third parties. We do not work with data brokers or advertising networks.

5. Google API Services Limited Use Disclosure

Draftroom's use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements.

Specifically:

• We only use Google data to provide and improve Draftroom's email drafting functionality.
• We do not transfer Google data to third parties except as necessary to provide the service (Anthropic Claude for draft generation), with user consent, or as required by law.
• We do not use Google data for advertising or to serve ads.
• We do not allow humans to read your Google data unless you give explicit consent, it is necessary for security purposes, to comply with law, or the data is aggregated and anonymized for internal operations.

6. Data security

We take the following measures to protect your data:

• All data is encrypted in transit using TLS 1.2 or higher.
• All data is encrypted at rest using AES-256 encryption via Supabase's managed PostgreSQL.
• OAuth refresh tokens are stored encrypted and are never exposed to client-side code.
• Row Level Security (RLS) is enabled on all database tables, ensuring users can only access their own data.
• API keys for third-party services are stored as server-side environment variables and are never bundled into client code.
• We use the principle of least privilege for all service integrations.

7. Data retention & deletion

We retain your data for as long as your account is active. When you delete your account, we revoke your Google OAuth tokens and perform a cascading deletion that removes:

• Your voice understanding and all contact voice profiles
• All drafts, draft runs, and associated metadata
• All learnings extracted from your editing behavior
• All Claude conversation exports
• Your Gmail connection and stored OAuth tokens
• Your reply patterns and contact reply profiles
• Your knowledge base items and integration connections
• Your inbox classifications and error logs
• Your user record and account information

Deletion is permanent and irreversible. Once initiated, all data is removed from our database within 30 days, including any backups. Your Stripe subscription is cancelled automatically upon account deletion.

If you want to delete your data without using the app, email help@withdraftroom.com from the email address associated with your account and we will process the deletion within 5 business days.

8. Your rights

Depending on where you live, you may have specific legal rights regarding your personal data. We honor these rights regardless of your location:

• Access & Export — you can request a copy of all data we hold about you. You can also export your data directly from your account settings.
• Correction — you can request that we correct inaccurate data.
• Deletion — you can delete your account and all associated data at any time from your settings, or by emailing us.
• Portability — you can request an export of your data in a machine-readable format.
• Restriction — you can ask us to stop processing your data while we resolve a concern.
• Objection — you can object to our processing of your data for specific purposes.
• Withdraw consent — you can revoke your Google OAuth connection at any time from your Google account settings or from within Draftroom.

For GDPR (EU/EEA residents): Our legal basis for processing your data is your explicit consent, provided when you connect your Google account and agree to our terms. You may withdraw this consent at any time. If you believe we have violated your rights under GDPR, you have the right to lodge a complaint with your local supervisory authority.

For CCPA (California residents): We do not sell your personal information. We do not share your personal information for cross-context behavioral advertising. You have the right to know what data we collect, to request deletion, and to non-discrimination for exercising your rights.

To exercise any of these rights, email help@withdraftroom.com. We will respond within 30 days.

9. Cookies & tracking

Draftroom uses only essential cookies required for authentication and session management. We do not use analytics cookies, advertising cookies, or tracking pixels. We do not participate in any ad networks or retargeting programs.

10. Children's privacy

Draftroom is not intended for use by anyone under the age of 16. We do not knowingly collect personal data from children. If you believe a child has provided us with personal data, please contact us at help@withdraftroom.com and we will promptly delete it.

11. Changes to this policy

We may update this policy from time to time. When we make material changes, we will notify you by email at least 14 days before the changes take effect. The “Last updated” date at the top of this page reflects the most recent revision.

12. Contact

If you have questions about this privacy policy or how we handle your data, reach out to us:

• Email: help@withdraftroom.com
• General support: help@withdraftroom.com

We read every message and will get back to you within 2 business days.