Watch your
sort itself.

Sieve clears the unwanted email that steals hours from your week—without getting between you and the people who matter.

Run free inbox audit
Before Sieve
GmailSearch mail
  • ☆Apex RecruitingStaff Engineer role at Series B fintechSolicitor
  • ☆Linear2 unread notifications on Acme board
  • ☆Adrian @ GrowthloopQuick 15 min Tuesday?Solicitor
  • ☆Stripe SecurityVerify your account or it'll be suspendedMalicious
  • ☆MercuryGusto has charged your Mercury card
  • ☆Hartwell CapitalPE growth fund — quick intro?Solicitor
  • ☆FinanceMeeting Notes: Project Alpha
  • ☆EmilyNew NDAs are live in the Dropbox
After Sieve
GmailSearch mail
  • ☆Linear2 unread notifications on Acme board
  • ☆MercuryGusto has charged your Mercury card
  • ☆FinanceMeeting Notes: Project Alpha
  • ☆EmilyNew NDAs are live in the Dropbox

Every day you sort through email you never wanted. Just to find the few that matter. Sieve does the sorting, so you don't.

Phishing and malicious mail get caught too.

Why Sieve

Get five hours back. Every month.

Sieve catches the clutter. The people and messages that matter come through clear.

Run free inbox audit
Setup

A couple minutes is all it takes.

Sieve Sign in to a cleaner inbox
Continue with Google
Continue with Outlook
What should Sieve filter?
Sales pitches Recruiters Cold outreach Newsletters
Keep intros from people I know.
Corrections sharpen it from there.
Maya Chen Re: launch checklist Inbox 3:42p
Adrian @ Growthloop Quick 15 min Tuesday? Solicitor 3:28p
Linear 3 new threads on Acme board Inbox 2:58p
Apex Recruiting Staff Engineer role Solicitor 2:51p
account-alerts Action required: verify now Malicious 12:09p
How it works

Your inbox arrives already clean.

Every email is checked against a series of various key signals.

  • Sender relationship Anyone you've emailed or replied to is trusted and skips the queue. Their mail heads straight to your inbox, with just a quick check of its links against known threats.
  • Authentication Sieve confirms the sender's domain actually sent the message. Forged and spoofed addresses are caught right here.
  • Threat signals Every link and attachment is checked against live threat feeds. Known-bad URLs and malicious files are stopped before they ever reach you.
  • Your preferences You handle the grey areas. A short note on what counts as noise to you, plus the mail you keep or archive, sharpens every call Sieve makes.
  • Labeled Caught mail gets an inbox label, Solicitor or Malicious, and leaves your inbox. Everything else passes through untouched.
Smart Unsubscribe

Stop the lists that keep finding you.

Sieve groups recurring list mail, shows why it is there, and lets you stop and file future mail or keep the sender from one calm queue.

  • Keep the few lists worth reading.
  • Stop unwanted lists and file future mail.
  • Protect known senders and real threads.
Smart Unsubscribe Review queue
Growthloop Weekly 6 this month · outbound list Stop & file
Founder Notes Kept · trusted sender Keep
Vendor webinar series Stopped · future mail filed Stop & file
Keep and block choices can be changed anytime.
Outreach Source Removal Available now

Reduce the databases that feed the outreach.

Outreach Source Removal finds sales-data sources, prepares removal requests, and tracks follow-up before that outreach reaches your inbox.

Removal workflow Live
Broker source Found
Request package Prepared
Follow-up record Tracked

Private and secure

Audited

Reviewed controls and access boundaries.

  • Google security review completed
  • SOC 2 Type II in progress with U.S.-based peer-reviewed auditors
  • Restricted Gmail access reviewed

Encrypted

Data protected at rest and in transit.

  • AES-256 encryption at rest
  • TLS for data in transit
  • Hosted in US-based infrastructure

Controlled Access

Sieve uses Gmail or Outlook access to classify and file messages. It sends only the Source Removal requests you authorize.

  • Restricted Gmail access reviewed
  • Internal safeguards constrain email actions
  • Disconnect and delete your Sieve data anytime
Read the full privacy explainer
Pricing

Built for your inbox, or your team's.

Basic
$10102 / month/ year $120
You save $18/yr

Filter what comes in

  • Filters unwanted outreach and learns from your corrections.
  • Review every decision in your dashboard.
  • One Gmail or Outlook inbox.
  • One-click Smart Unsubscribe from mailing lists.
  • Basic Deep Clean: Bulk Sweep old newsletters and recurring clutter.
Start free trial now

7 days free. Cancel anytime.7 days free. Cancel anytime.

Pro
$20204 / month/ year $240
You save $36/yr

Filter it, and make it stop

  • Everything in Basic.
  • Second inbox included (Gmail or Outlook).
  • Pro Deep Clean: Smart Scan cold outreach and one-off clutter.
  • Outreach Source Removal.
  • Breach monitoring.
  • Priority support.
Start free trial now

7 days free. Cancel anytime.7 days free. Cancel anytime.

Team
$15153 / seat / month/ seat / year $180
You save $27 per seat/yr

Filter it for everyone

  • Everything in Pro, for every member.
  • One admin dashboard for the whole team.
  • Seats and billing managed in one place.
  • Organization-wide filtering rules.
  • Shared trusted senders and threat signals.
Start free team trial now

14 days free. Cancel anytime.14 days free. Cancel anytime.

FAQ

Common questions

Can I tell Sieve what to filter and what to keep?
Yes. You can write a short plain-English description of your grey zone, such as "I'm not interested in recruiter pitches," and add specific senders or domains to your trusted list (like your accountant). It's guidance for the classifier, not a scriptable rules engine. The fastest correction is just dragging a misclassified message back to your inbox; that's recorded automatically and future classifications improve.
What happens if the classifier gets it wrong?
Drag the message back to your inbox or remove the label. The correction is recorded automatically and future classifications improve. You can also add senders to your trusted list.
Will the AI read mail from my coworkers or people I know?
No. Your own company domain, trusted senders, and anyone already in a thread with you are cleared by deterministic rules before the model runs. The classifier only sees mail from senders you have no relationship with.
Do you read my email content?
Only when a stranger's message needs AI classification: the subject line and a short plaintext excerpt of the body (about 2,000 characters) are sent to the AI for a single classification. It is never used to train any model, and is deleted after 30 days. Your contact list never leaves our servers: name-drops are checked against it locally, and only the result is shared with the model. We store no message content ourselves, except the subject line of mail we flag, which is encrypted and deleted after 30 days; otherwise only metadata like the label and which signal drove it.
What Gmail or Outlook data do you access?
For Gmail, we request gmail.modify to read, classify, label, and archive mail. For Outlook, we request Microsoft Graph Mail.ReadWrite for the same filing actions and Mail.Send only for Source Removal requests you authorize. We check sent history to recognize people you already email. We inspect attachment names and types but never store attachment files. Provider refresh tokens are encrypted at rest.
Can I see what was filtered?
Yes. Your account shows every classification from the last 90 days, including which signal drove the decision.
What happens when I cancel?
Canceling stops your subscription: filtering pauses when your subscription ends, and your inbox is left untouched from then on. Re-subscribe anytime to turn it back on. Canceling does not delete your data or disconnect Gmail or Outlook. Only Delete account, in Settings, erases the data we store; it also disconnects your mailbox and cancels billing. Existing provider labels, categories, or folders stay until you remove them.