SHOO, SPAM. NOT TODAY.
A determined little guardian for unwanted mail. Spots learned spam patterns, gives a playful “shoo!” and parks the message in AI SPAM.
Try it free on Windows ↓30 days free. Then US$45/year. Optional packs sold separately.
Your regulars are welcome.
People you have emailed belong in your Inbox. ANIMAIL protects the exact addresses in your human To, Cc and Bcc correspondence across connected accounts, alongside your saved contacts and trusted exceptions. This protection applies at every position of the Low–Strong slider. Receiving an email, seeing an address in autocomplete or sending an automatic response does not turn a stranger into a trusted correspondent.
A hello worth remembering.
Customer protection remembers addresses independently of cached Sent emails and delivery records. Cleaning those up or restarting the app keeps the protection ANIMAIL has learned. The update also learns from all retained accepted sending history and cached Sent mail, including correspondence beyond the most recent 2,000 deliveries. If someone’s entire earlier history was permanently deleted before it could be learned, add their address to Contacts or trusted exceptions.
A final check before a reply.
If you email a previously flagged person, ANIMAIL removes the automatic ban and cancels pending or connecting spam replies to that address. It checks protection, contacts and reply settings again after connection preparation and authentication, just before submitting to Gmail, Microsoft or SMTP. Past detection history remains available: use Not spam · restore & trust to restore an older mistaken detection. Replies already handed to your provider cannot be recalled by this protection.
A guardian with a little character.
Enable AI Spam Guardian to check new incoming Inbox mail with an offline machine-learning model. Strong detections bring a playful AI Detected Spam robot notice, move the message into the local AI SPAM view and add the exact sender to ANIMAIL’s ban list.
Your mail stays on your PC.
No cloud AI account or API key is needed. The model reads bounded message text locally; it does not upload your mailbox or scan attachments. AI SPAM combines detections from your receiving accounts, with complete messages and attachments retained. Provider originals stay available: quarantine and bans apply inside ANIMAIL on this PC.
Your reply. Your choice.
Automatic replies are a separate, initially off setting. Write your own plain-text response and choose whether to enable it. Replies are queued only for eligible authenticated direct senders, with duplicate, loop, expiry and rate-limit safeguards. Lists, bounces, no-reply addresses and unverified senders are skipped. Recent detections show whether a reply was queued, accepted or skipped.
A second chance for a genuine hello.
AI can make mistakes. Its model scores are not measured accuracy percentages, and messages without a strong enough match stay in the Inbox. Not spam · restore & trust returns a mistaken detection, removes its local ban, trusts the sender and teaches the local model. Restoring a mistake also cancels an unsent automatic reply.
A little lesson, clearly saved.
Mark mail as spam to move it out of the Inbox and into AI SPAM. A friendly Teaching your AI animation confirms Lesson learned after the correction is saved. Manual corrections work even when automatic detection is off and never send a retroactive automatic reply. Reset learned corrections erases saved examples without deleting messages or quarantine.
Choose how strong your guardian should be.
A Low–Strong slider lets you quarantine only the strongest matches or catch more unwanted mail. New setups start gently; your existing saved level is preserved. Contacts, trusted exceptions and recipients of your own outgoing correspondence across connected accounts are protected at every strength. Queued automatic replies recheck these protections immediately before sending. AI can still make mistakes with unfamiliar senders.
Start gently. Stay in control.
Open Tools → AI Spam Guardian to choose receiving accounts and configure detection. Both detection and automatic replies start off. Existing downloaded mail and first-sync history do not trigger surprise replies. Preview the robot, turn its notice off independently, or use quiet hours and reduced motion. Settings, local learning, bans and quarantine travel in complete encrypted backups.
A LITTLE LESS ORDINARY.
Ready for a happier hello?
Get the native Windows app, with a 30-day trial of the core mail features.
