Organised by what you're trying to do, not by who wrote it
Most help centres bury the answer in a thousand articles, and the industry's own numbers say only one in seven people resolves their problem that way. This one is organised the other way around — by the situation you're in right now. Find your case below for the answer or the right next step, and when self-service genuinely isn't the fastest path, the operator team is a direct route rather than a ticket queue buried three clicks deep.
Where to start, depending on what you're trying to do
Most help centres are a search box and a thousand articles, and the industry's own numbers say only about one in seven people actually resolves their problem that way. The gap isn't that the answers are missing — it's that they're organised for the people who wrote them rather than the person who's stuck. This page is organised the other way around: by what you're trying to do right now. Find the situation that matches yours, and the answer or the right next step is there. When self-service genuinely isn't the fastest path — and for infrastructure problems it often isn't — the operator team is a direct route, not a ticket queue, and we say so plainly rather than hiding the contact option three clicks deep.
One structural note, because it increasingly matters: this page is written to be read by people and by the AI assistants people now ask for help. That means plain questions with direct answers rather than marketing prose, so that whether a human scans it or a model retrieves it, the answer comes back accurate.
Getting started — your first send
What do I need before I can send? Administrator access to your domain's DNS, so SPF, DKIM and DMARC records can be added; the sending domain or subdomain you intend to use; and a decision on whether your application will connect over SMTP or the API. For managed plans, the operator team handles the DNS and authentication setup with you rather than leaving you to it.
SMTP or API — which should I use? If your application already speaks SMTP, the relay is a configuration change: host, port, username, password, and you're sending. If you'd rather the sending logic live in product code with structured responses and webhooks, the API fits better. Neither is more "advanced" — they're different integration shapes, and the comparison of sending layers walks through the trade-off in detail.
How long until I'm actually delivering? SMTP relay can be live within the hour once DNS propagates. The honest constraint isn't setup speed — it's reputation. A brand-new sending domain and IP have no reputation, and sending at full volume immediately is the fastest way to land in spam. Where dedicated IPs are involved, the warming schedule is the real timeline, and it runs over weeks, not hours.
Authentication — SPF, DKIM and DMARC
This is the layer where most deliverability silently dies, so it's worth getting exactly right. Every check below can be run on any domain using the free tools on this site, including ours.
My SPF record stopped working when I added a service. SPF has a hard limit of ten DNS lookups, and each include: can chain to more. Add enough services and you blow past the limit, at which point the whole record can fail. The fix is flattening — resolving the includes to their IP ranges. The SPF flattener does this and warns you when a source record is dangerously large.
How do I check whether DKIM is set up correctly? The DKIM tester retrieves the public key at a given selector, estimates the key length, and flags the common misconfigurations — a 1024-bit key where 2048 is expected, a malformed record, a selector that doesn't resolve.
What does p=none actually do, and is it enough? A DMARC policy of p=none monitors but enforces nothing — it tells you who's sending as your domain without stopping anyone. For protection, and for the bulk-sender requirements that Gmail and Yahoo now apply, you need to move toward p=quarantine and then p=reject. The DMARC checker shows your current policy and reads it back in plain language.
Deliverability — when mail isn't landing
My email is "delivered" but recipients say they never got it. Delivered and inboxed are different things. "Delivered" means the receiving server accepted the message; it can still be filed in spam. This gap is the single most common deliverability confusion. Check authentication first, then sender reputation, then content and list hygiene. The status page explains why we treat inbox placement as the metric that actually matters.
One of my IPs got blocklisted. What now? First, identify which list — the blocklist lookup checks the major DNSBLs at once. The delisting process and the right response depend heavily on which list and why. On managed plans this is exactly the kind of incident the operator on call handles directly, usually having dealt with the same listing recently on other infrastructure.
My delivery rate dropped suddenly with no config change. Sudden drops with no change on your side usually trace to one of a few causes: a reputation shift from a content or volume change, a receiving provider tightening filters, or a blocklisting you haven't spotted yet. The diagnostic order matters, and it's faster to talk it through with someone who can see the sending data than to guess. Reach the operator team.
Migration — moving from another provider
Will I have to change my code? If you're on SMTP, usually just configuration — host, credentials, done. If you're on another provider's API, it's a field-mapping exercise, since REST conventions are similar but endpoint paths and field names differ. We've documented the specific path from each major provider: SendGrid, Mailgun, Postmark, Amazon SES and others.
Will migrating wreck my deliverability? It can, if done carelessly, because a new IP starts cold. Done properly, you carry your authentication forward, warm the new dedicated IP gradually, and run a dual-send overlap on critical traffic so password resets and billing never go dark during the cutover. The migration guides lay out the typical four-week sequence.
Can you take my suppression list? Yes. Hard bounces, spam complaints and unsubscribes export from your current provider and import into ours, which is essential for reputation continuity — a recipient who unsubscribed shouldn't suddenly start receiving mail again because you switched infrastructure.
Account, billing and compliance
Where's the data processed, and what about GDPR? EU-only by default — Sweden and Germany — which removes the international-transfer paperwork entirely. The Data Processing Agreement is the binding document for your legal team, and the trust page maps the GDPR, NIS2, DORA and ISO 27001 obligations honestly.
How does billing work? Flat monthly pricing in EUR with the message volume, dedicated IPs and support included — no per-IP add-ons, no surprise overage at month end. The pricing page lists every tier.
When self-service isn't the fastest path
For a misconfigured record or a "which layer do I need" question, the answers above and the free tools resolve it quickly. For a live deliverability incident, a blocklisting, or anything where someone needs to see your sending data to help, skip the search box — the operator team responds directly, and on managed plans it's the same people who run the infrastructure, not a tier-one script.
The things people ask that don't fit a category
Do the free tools send my data anywhere? No. The DNS-based tools on this site run the lookups in your browser against public DNS over HTTPS; the query and result don't touch our servers. You can verify this in your browser's network tab.
Can I use these tools on a domain that isn't a customer? Yes. The DMARC checker, SPF flattener, DKIM tester, blocklist lookup, TLS report decoder and BIMI validator work on any domain. They're published precisely so you can audit anyone's setup, including ours before you become a customer.
What if my question is none of these? Then it's exactly the kind of question worth asking a person. Email infrastructure has long tails of edge cases that no help centre fully covers, and the value of a small named team is that the answer comes from someone who has seen the edge case before. Ask directly.