Frequently Asked Questions

Email Ferret is a Gmail filter for AI-generated cold outreach. It reads incoming mail, scores how likely each message is to be an automated sales pitch, and labels the ones that are so your inbox holds the mail that was actually written for you. It runs in the background on Gmail and Google Workspace accounts, and there is nothing to install on your computer.

Setup is a Google sign-in and takes about a minute. You connect a Gmail account, Email Ferret creates its labels, and filtering starts on the next message that arrives. There are no rules to write, no keyword lists to maintain, and no training period you have to sit through before it starts working.

Email Ferret scores mail as it arrives, so the immediate effect is on new messages. That means you see a clean inbox going forward rather than a one-time cleanup of a backlog. If your goal is to dig out from thousands of old cold emails, use a Gmail search and bulk actions for the backlog, then let Email Ferret keep the inbox clear from that point on.

Most people see flagged messages within the first day, because cold outreach arrives constantly. The clearer signal comes after about a week, once enough mail has passed through to show the pattern: the sales sequences land under their label and the inbox holds conversations, receipts, and mail from people you actually know.

No. Filtering happens on the server, so it works whether you read mail in the Gmail web interface, on your phone, or in a desktop client. The Chrome extension is a convenience layer that surfaces scoring and lets you allowlist or block a sender without leaving Gmail.

Gmail's spam filter is built to catch mail that is malicious or fraudulent — phishing, malware, and bulk junk from bad domains. AI-generated cold outreach is none of those things. It passes SPF, DKIM, and DMARC, it comes from a real company on a warmed-up domain, and it is written well enough to read like a human wrote it. Gmail delivers it because by its standards the message is legitimate. Email Ferret filters on a different question: was this written for you, or generated and sent to a list?

It scores behavior rather than reputation. The signals include whether the sender is new with no prior conversation history, whether the message uses sales framing and a call to action, whether it looks like one step in a timed follow-up cadence, whether the personalization is generic — your name dropped into a template — and whether it carries fingerprints of known outreach automation. Messages that score high enough go to a language model for a second check before being labeled.

Blocklists work on identity, and cold outreach identity is disposable. A sending platform can rotate through hundreds of fresh domains, warm each one for weeks, and retire it before enough people report it. By the time a domain is blocked it has been replaced. Behavior is harder to change, because the thing that makes cold outreach effective — sending at volume to strangers on a cadence — is exactly what makes it detectable.

Remove the label and add the sender to your allowlist, and mail from that address always reaches your inbox from then on. Because nothing is deleted, a mistake costs you the seconds it takes to find the message under its label. The allowlist is the right tool for the predictable cases — your recruiter, your vendor, an investor you do want to hear from cold.

Flag it, and if the sender is a repeat offender add them or their domain to your blocklist so anything from them is filtered regardless of score. Some messages are genuinely ambiguous — a founder writing to you personally about something relevant looks a lot like a good cold email, because it is one. Email Ferret is deliberately tuned to let those through rather than risk burying real mail.

Yes. Every scored message shows the signals that contributed and the score they produced, so a flag is something you can inspect and disagree with rather than a verdict you have to take on faith. If the reasoning looks wrong, that is the point at which allowlisting or blocking the sender is a considered decision.

No. Message bodies are analyzed in memory and discarded — they are never written to a database. What persists is the minimum needed to run the interface you use to review and correct decisions: subject lines and message ids, encrypted at rest with AES-256-GCM and isolated per user with row-level security so no account can read another account’s data.

No. Email Ferret has no advertising business, no data brokerage, and no third party that receives your mail for its own purposes. The business model is the subscription — you pay $5 a month, which is what makes it possible to not monetize your inbox some other way.

Only what filtering requires: permission to read incoming messages so they can be scored, and permission to apply labels and archive so decisions can be acted on. Access is granted through Google OAuth, so your Gmail password is never given to Email Ferret and is never stored. OAuth tokens are encrypted before they are persisted.

Yes, at any time, and from either side. Disconnect the account inside Email Ferret, or revoke the app from your Google account's security settings. Filtering stops immediately. Labels already applied stay in Gmail, because they are ordinary Gmail labels — you can delete them yourself if you want no trace left.

Cancelling ends the subscription and stops filtering. You can delete your account outright, which removes the stored subject lines, message ids, allowlist, blocklist, and OAuth tokens. Since message bodies were never stored, there is no archive of your mail to delete — it was discarded at the moment it was analyzed.

Five dollars per month per mailbox. There is one tier, so the price does not change based on how much mail you receive or which features you use. Connecting a second Gmail account — a work inbox alongside a personal one — is a second mailbox.

Yes: fourteen days with full access, and no credit card to start. The trial is deliberately long enough to cover a couple of full weeks of inbound mail, which is what it takes to judge a filter — a day or two tells you it labels things, two weeks tells you whether the labels were right.

Nothing is charged automatically, because there is no card on file unless you added one. Filtering stops when the trial ends and resumes when you subscribe. Your allowlist, blocklist, and settings are kept, so subscribing later picks up where the trial left off rather than starting over.

Yes, from the billing page, with no cancellation process to sit through and no retention call. Access continues to the end of the period you have already paid for.

Yes. Teams get the same filtering plus a shared allowlist and blocklist, so one person marking a sending domain as unwanted covers everyone rather than each colleague blocking it separately. Get in touch through the contact page for team and enterprise pricing.

Not currently — Email Ferret supports Gmail and Google Workspace only. That is a deliberate trade: several of the strongest detection signals depend on conversation history and thread state that Gmail exposes well, and supporting one provider properly beats supporting four shallowly. If you use Outlook, this is not the tool for you today.

Yes. Connect as many as you like — each is billed as its own mailbox and keeps its own allowlist, blocklist, and folder rules. That separation matters in practice, because a vendor you want to hear from at work is often one you do not want in your personal inbox.

No. The default action is to label, which leaves every message exactly where Gmail put it and searchable as normal. You can additionally choose to archive flagged mail so it skips the inbox, but archiving is not deletion — the message stays in All Mail. Nothing Email Ferret does is irreversible.

Yes. You can rename the labels, create custom folders for categories that matter to you, and write routing rules that send particular senders or subjects to particular folders. If you already have a Gmail labeling scheme you like, Email Ferret can file into it instead of imposing its own.

Yes, and it is worth knowing the order. Your own Gmail filters run first and win — anything you have explicitly told Gmail to do with a message still happens. Email Ferret scores what remains. So a filter that stars mail from your domain keeps working exactly as before.

No. Scoring happens on our servers after Gmail receives the message, not in your browser, so Gmail itself is untouched — same speed, same interface, on every device. The only visible change is that some messages arrive already labeled.