Skip to content

Security

The strongest security control is not holding the data at all.

Last updated: 2026-10-11

At a glance

  • Google OAuth verification approved April 21, 2026, and CASA Tier 2 assessed by TAC Security.
  • We never fetch or store full email contents.
  • If you allow AI, some email details go to Anthropic to write text. See “AI processing” below.
  • Deleting your account: in 7 days, or right away if you choose.

What we store, and what we never store

DeclutrMail’s promise is literal: We never fetch or store full email contents. DeclutrMail requests only the Gmail details listed below.

Gmail details DeclutrMail stores:

  • Gmail message and conversation IDs
  • Sender name and email address
  • Subject line
  • Gmail preview snippet (the short text shown in your inbox list)
  • Date received
  • Gmail labels
  • Your Gmail label names
  • Read or unread state
  • Whether a message was sent by you
  • Recipient email addresses from To and Cc on email you sent
  • Unsubscribe links and whether one-click unsubscribe is supported
  • Estimated Gmail message size

DeclutrMail never fetches or stores:

  • Full email contents
  • Email HTML
  • Attachments
  • Images embedded in emails
  • Raw email source
  • Email headers other than From, Subject, To, Cc, and unsubscribe information

We don’t fetch or store full email contents or attachments, so there’s no message body in our systems to leak. Subject lines and Gmail preview snippets can still contain sensitive information, so we list and protect them explicitly.

The Gmail permission we request, and why

DeclutrMail requests one Gmail permission, gmail.modify, plus openid and your email address to identify the connected account. DeclutrMail’s job is to change your email when you tell it to: archive, label, delete and unsubscribe. gmail.modify is the permission that allows those changes.

This permission is broader than what DeclutrMail uses. The app requests only the listed fields and never asks Gmail for the full or raw email formats that include complete contents and attachments. You can revoke access at any time from Settings or from your Google account permissions page.

Encryption

All data is encrypted in transit (TLS) and at rest. Your Gmail OAuth tokens get an extra layer: each token is envelope-encrypted with its own fresh 256-bit data key (AES-256-GCM), and that key is in turn wrapped by a key-management-service key that never enters the application process. Tokens are never sent to your browser and never included in data exports.

Google verification and independent assessment

Google OAuth verification approved April 21, 2026. It covers the single restricted permission we request, gmail.modify.

CASA Tier 2 assessed by TAC Security. Its letter of validation was submitted to Google on April 15, 2026. Apps that use restricted Gmail permissions must complete Google’s Cloud Application Security Assessment through an authorized lab.

Brand verification — approved August 18, 2026. Google’s consent screen shows DeclutrMail’s verified name.

Verification is recertified every year. Adding a permission or changing our consent screen needs a new review.

AI processing

DeclutrMail uses Anthropic, an AI company, in two places. Neither one receives full email contents or attachments.

Suggestion explanations, on every plan including Free: only after you allow it in the app, DeclutrMail can send Anthropic the sender’s name and email domain, the suggested choice and the rule behind it, how much they email you and how much of it is marked read, and their Gmail category, to write a one- or two-sentence reason for a suggestion. It never includes subject lines, preview snippets or email contents. Until you allow it, or if you turn it off, explanations are written from our rules without AI.

Daily Brief (part of Pro; waitlist only for new accounts): to write your summary, DeclutrMail can send Anthropic the sender, subject line and Gmail’s preview snippet of each email it covers, only after you allow it in the app. If you don’t allow it, you get a basic version written without AI.

You can turn AI off any time in Settings → AI. Explanations then go back to our rules, and the Daily Brief to its basic version.

We don’t use your email to train AI models. Anthropic sets its own retention and training terms: Anthropic states that standard API inputs and outputs are deleted within 30 days, except where longer retention is required for usage-policy enforcement or law, or different contract terms apply.

DeclutrMail’s use and transfer of information received from Google APIs follows the Google API Services User Data Policy, including the Limited Use requirements.

Exactly what is sent, and when: Privacy Policy, AI processing.

No ML category prediction

DeclutrMail does not use machine learning to predict email categories or route senders. It can automatically protect a sender using fixed product rules when strong engagement signals, such as your reply history, cross the documented threshold; you can review and change that protection. Mail-changing automation follows preset rules you explicitly enable — never a model’s guess. We also do not use Gmail data to train generalized AI or machine-learning models.

Leaving cleanly

You can disconnect an inbox (which revokes our Google access, stops syncing, and preserves its historical DeclutrMail record for reconnection), delete one inbox’s saved data, or schedule deletion of your whole account from Settings. Account deletion waits at least 7 days, and also waits for any open undo window, so undo keeps working for its full window. Undo windows run 30 days on every plan, so a recent action commonly puts deletion up to 30 days out; you can waive both with a typed confirmation to delete immediately. Details are in the Privacy Policy.

Report a vulnerability

If you believe you’ve found a security vulnerability in DeclutrMail, email privacy@declutrmail.com with the details. We review every report and will respond, and we ask that you give us reasonable time to fix an issue before disclosing it publicly.