SlyReply
PlaygroundCatalogDocs
Log in Get Started

We use a strictly-necessary session cookie to keep you logged in. With your permission we also load privacy-friendly analytics (Plausible — no cross-site tracking) to see how the site is used. Analytics stays off unless you accept. Cookie Policy

Reference

Fair Use & Limits

How SlyReply's fair-use policy works — a short cooldown after a busy burst, and the generous monthly limits behind it.

Fair Use & Limits

SlyReply is flat-rate — one price a month, no per-message charge. A fair-use policy sits behind that price so a single account can't run up unbounded cost. It is deliberately generous: normal day-to-day emailing never reaches it.

A short rest after a busy burst

If an account sends a large burst of replies in a short window it takes an automatic 3-hour cooldown. Paid plans get a larger short-burst allowance than the free plan — roughly 20 replies within 60 minutes on Free, ~100 on Pro, and ~300 on Power before the cooldown kicks in, so a busier paid account isn't held to the free rate. (These are the production defaults; non-production environments like preview / sandbox tighten the thresholds so the cooldown can be exercised in testing.) During the cooldown:

  • Every inbound email is held and re-processed automatically the moment the cooldown ends, so in normal cases nothing is lost.
  • The first email from any given sender during the cooldown gets a friendly, one-time notice — "agent is taking a short break; your email is in the queue, no need to resend, expect a real reply around HH:MM." Further emails from the same sender during the same cooldown are also held but don't get a repeat notice (the first one already covers them — that's deliberate, to avoid burying the sender in resting messages).
  • If a very large volume backs up during the cooldown, the oldest held emails get the same "short break" note instead of being queued. The threshold sits well above any normal busy hour.

The cooldown clears itself; there is nothing to buy and nothing to do. In practice it's almost always a sign of a misconfiguration — an agent caught in a mail loop, or wired behind a busy web form — rather than genuine use.

Generous monthly limits

Behind the cooldown, each plan has a generous monthly allowance of replies and images. It's an outer backstop, not a meter:

PlanReplies / monthImages / month
Free50— (no image generation)
Pro2,00025
Power10,000150

These reset every 30 days. They sit far above what a typical user sends, so reaching one is rare — and if you get close, your profile says so plainly, well before.

Seeing where you stand

Your Usage & Billing page shows a calm, qualitative status for the month — quiet, steady, busy, or resting — alongside the plain count of replies used and the date your period rolls over. When an agent is resting, the page tells you when it's back. The intent is for the page to be glanceable rather than something to budget against in detail; the monthly ceiling sits far above normal use, so the qualitative status is what matters day-to-day.

What counts as heavier use

Fair use is measured in replies and images, not raw tokens — but behind the scenes a reply's size does vary, and a few habits use the generous monthly allowance faster than others. Roughly, heavier:

  • Long emails and long threads. The agent re-reads the whole conversation each turn, so a 30-message thread is a bigger call than a fresh one-line question.
  • Attachments. A large PDF or image is a lot for the model to take in at once (Pro and Power only).
  • Image generation. Each image counts against the separate monthly image allowance — Pro 25, Power 150.
  • The most capable models. A GPT-5-class reply (Power) is a heavier call than a fast-model reply.

And lighter:

  • Short, focused questions with a clear ask.
  • Fresh threads for new topics — starting a new conversation drops the old history the agent would otherwise re-read every reply.
  • The fast models (Haiku / GPT-class minis) for everyday back-and-forth.

None of this is worth micro-managing — the monthly limits sit far above normal use. It only matters if an agent is unusually busy, e.g. wired behind a high-traffic inbox or a public form.

Abuse protection

Separately, SlyReply drops obviously abusive inbound mail immediately — mail loops, and floods of mail from many different senders at once — to protect both your account and the people emailing it. Ordinary correspondence is never affected.


Related: Plans & Pricing · No Surprise Bills

← Previous
Plans & Pricing
Reference
Next →
Conversation & Attachment Limits
Reference
SlyReply·Email an AI. Get a reply.
DocsCatalogPricingContactTermsPrivacyDPACookiesSecurity

SlyReply is a service of Sherman Studio Ltd, a company registered in England and Wales (company number 17244879).

Registered office: 71-75 Shelton Street, Covent Garden, London, WC2H 9JQ.

© 2026 Sherman Studio Ltd. All rights reserved.