Questions, answered plainly

No. You sign up, get an API key, and send. We create and operate everything underneath — on paid plans that means real infrastructure dedicated to you alone, with its own reputation, quota, and suppression list. You never see a console, a template, or a review form.

Your environment is created without root credentials (nobody has a master password, including us), restricted by policy to email services, and reached by our control plane only through short-lived, individually locked credentials. No standing keys exist anywhere in the system.

Through volume plans, like every sender you've used: Pro at $20 includes 50,000 a month, Scale at $90 includes 600,000, Growth at $299 includes 2,000,000 — and past the bundle it is $0.15 per 1,000, published on the pricing page, the same number the calculator uses.

The rate is ours, not a pass-through. Raw sending costs are public knowledge; the difference funds what dedicated infrastructure costs to run — provisioning your environment, clearing the production review, warm-up guardrails, quota raises, and the transfer concierge if you ever leave. Every bundle stays above our real cost, enforced by a test in CI, so no plan is a loss-leader we'll claw back later.

The pricing reads the same on purpose. The difference is underneath: they pool customers on shared infrastructure, so a stranger's bad list can become your deliverability problem. Here, every paid plan runs on infrastructure that carries you alone — your reputation record, your quota, your standing, physically separate from every other customer's.

If you send 2,000 emails a month, take Resend's free tier — we'd rather say that than sell you something you don't need. The case for us starts when deliverability is revenue: when you or your clients cannot afford to share a reputation with strangers.

One isolated sending identity inside your dedicated environment: its own suppression namespace, its own reputation policy, and any number of sending domains. It is the smallest thing that can be paused without affecting anything else.

Brands are unlimited and never billed, on any plan. If we charged per brand, you would merge two clients onto one sending identity to save money — and keeping client reputations apart is the entire point of the product.

Yes, and in writing: transfer is a contractual right, not a favor. Your dedicated environment transfers into your hands and becomes fully yours — same identity, same verified domains, same DKIM keys, same earned reputation — along with a full export of your suppression list and message history.

Honest expectations: plan on days, not minutes. The hand-over includes steps only you can perform — adding your own payment method, verifying contact details — and quotas may be re-reviewed on the way out. The runbook is documented and we walk it with you. No shared-pool sender can offer any of this, because there is nothing of yours to hand over.

Software, through one narrow door: our control plane obtains short-lived credentials scoped to email services through a single, individually locked entry point — sending, identities, suppression, and the event streams. There is no credential column in our database, encrypted or otherwise.

Humans, almost never: environments are created with no root credentials at all, the one place that could mint access is workload-free behind hardware MFA, and every environment is policy-restricted to email services. We test our own locks by trying the door without the key — and refusing the setup if it opens.

Two walls. In the cloud: one isolated environment per paying customer — the hardest boundary money buys. Reputation, quota, and standing are physically separate resources, not rows we filter.

In our database: PostgreSQL row-level security keyed on the organization, with the application connecting as a non-owner role so the policies actually constrain it. That is deliberately not the same thing as filtering in application code — this is the layer that still holds when someone forgets a WHERE clause.

From our database: your organization and brand structure, sending domains, message metadata (recipients, subjects, statuses, events), and suppression lists — but no message bodies or attachments, and no stored credentials, because none exist to steal.

The deeper question is the fleet itself, and we answer it structurally rather than with adjectives: environments hold no root credentials, are policy-locked to email services, and sit behind a workload-free management layer with hardware MFA. The trust page describes the whole model plainly — it is the question a security reviewer is actually asking, so we answer it in public.

Yes. Pause is a single API call or support message, effective immediately, no questions asked — your API keys are yours to revoke, per brand or entirely.

And the deeper version of the same right: the transfer. Your account, your domains, your reputation, and your data can leave our custody entirely, runbook documented. We never hold an account hostage over anything except an unpaid bill, and the terms say so.

Not directly, and we would rather be blunt about it than imply otherwise. We do not warm IPs into Gmail, we have no relationships with mailbox providers, and we cannot escalate on your behalf. Nobody selling email infrastructure honestly can promise the inbox.

What we own is your standing upstream — the production review, quota raises, any case that gets opened — and the structure that keeps infrastructure from ever being why you missed the inbox: a dedicated environment so nobody's mistakes touch you, per-brand reputation isolation so your own clients can't hurt each other, correct DKIM and DMARC on every domain, and a suppression list that actually stops repeat sends.

We pause that brand at 2% bounce or 0.05% complaint, measured over a rolling window. Upstream reviews trigger at roughly 5% and 0.1%, so we act well before your standing is at risk — and we pause the brand, not the environment, so your other brands keep sending.

You get the numbers that triggered it, not just a status change. These guardrails are also what keep your environment's standing clean — which is what keeps quota raises fast and reviews boring.

Yes — attached to your environment and the right brand, billed on our invoice at cost: $24.95 per month per IP, labeled as what it is.

Worth knowing before you ask: managed dedicated IPs spill onto a shared pool during warm-up, and IP pools are capped at 50 per region, which is not adjustable.

Yes, and pretending otherwise would be lying. If the platform underneath us ever pauses sending, the case is ours to open and argue — that is part of what the plan pays for — and our early-pause guardrails exist precisely to keep you far from that line.

The structural honesty: every dedicated environment sits inside our fleet, so we run its aggregate standing as an asset — strict thresholds, paced onboarding, hard isolation between environments. One bad actor cannot become everyone's problem, and keeping it that way is the operational job we charge for.

Minutes to first send: sign up, verify your domain (four DNS records — DKIM is a single record under your own name, and each row reports its live DNS state so a typo is named, not guessed at), and send from the labeled warm-up pool immediately. Verification re-checks itself; you never sit refreshing a screen.

Your dedicated environment arrives with the first paid plan: we provision it, submit the production-access review with your real use case, and flip your domains over the moment it clears — typically one to three days. It is a human review with no SLA, and we show you its status rather than hiding the wait.

A set of our own sending environments, kept deliberately clean, that carries Free-plan sending and the first days of a paid plan while your dedicated environment clears review. It is labeled in the API — routed_from tells you exactly what carried every send — and hard-capped: 3,000 a month with ramped daily limits on Free.

It exists so that day one is instant without ever being a shared-pool product in disguise: the pool is a bridge with your name on the far end, not the destination.

No. Your dedicated environment hosts unlimited brands, each fully isolated inside it — a holding company typically runs thirty brands on one environment.

Agencies flip that: one dedicated environment per client (so client reputations never share a roof), each holding that client's brands. Environments beyond your plan's bundle are $12–15 a month; resolution is most-specific-wins — domain, then brand, then your default.

Not yet. Today this is a REST API and a CLI, and the launch pricing reflects that. The dashboard is what we are building next.

If a UI is a hard requirement for you right now, we would rather you waited than paid for something that is not there.

Yes — over the API today. Claim support@ (or a catch-all) on any verified domain, point one MX record at us — the mailbox call returns the exact record — and mail lands in a real inbox: parsed, sanitized, full-text searchable, with replies correlated to the sends they answer. Mail to unclaimed addresses on your domains is quarantined, never dropped, and the raw message is always kept.

The webmail app for humans is in active development. The API is live now — an AI agent can already work a support inbox end to end.

Our commitment is transfer-on-shutdown: the same documented runbook that lets any customer take their account applies to every customer before the lights go out — your account, domains, DKIM keys, reputation, suppression list, and history leave with you.

That is the difference between an account held for you and a pool you were renting a slice of: there is something real to hand over.