Resend's simulated delivery, bounce and complaint recipients help test application event handling. They do not establish where Gmail, Outlook or another real mailbox places an original message. Use simulations for the application behavior they exercise, then a separate permitted real send through your actual route for LDM mailbox observation.
Know what the simulated recipient represents
Resend documents test addresses including delivered@resend.dev, bounced@resend.dev and complained@resend.dev. These produce predefined event scenarios. A simulated delivered result is useful for checking how your application records a success; it is not a measurement of a customer's primary Inbox.
Keep simulated complaints separate from mailbox-folder observations. An application receiving a complaint event and a copy observed in Spam are different facts. Do not count test-address results in a seed panel's received denominator or report a simulation as a real provider benchmark.
Exercise application handling deliberately
Choose the event scenarios relevant to your application. Check that a notification reaches the expected consumer, maps to the correct job and produces the intended internal state. Use non-sensitive job fixtures and distinguish a new event from a repeated delivery of the same notification.
Record the scenario and evidence precisely: the selected test recipient, service message reference, application job and processing outcome. A label such as email test passed is too vague to reveal whether it refers to payload construction, event consumption or an actual observed mailbox location.
Prepare a representative real message
Select the real application route you need to examine, including sender identity, renderer, template version and submission implementation. Preserve the final subject, HTML, text alternative, links and required unsubscribe behavior appropriate to the message type.
Use harmless fixture data. Do not send usable password-reset tokens, private customer information or valid sensitive action links to test mailboxes. A representative structure can be exercised with safe values without pretending that customer-specific authorization behavior has been validated.
Confirm that your Resend account, sender configuration and permissions allow the intended external recipients. A test address on the provider's own domain is not evidence that an arbitrary external recipient is permitted. Follow the current supported account and domain setup rather than attempting to evade a restriction.
Observe the actual route
Create an LDM test for the sender and use the current recipients supplied for that run. Generate fresh originals through the intended application path. Do not forward a simulated or personal copy, since forwarding introduces another route and message context.
Retain application submission evidence, service identifiers and timestamps. Resend's current Message-ID guidance documents identifiers useful for cross-system correlation. Match the available reference with the right recipient and observation, rather than relying only on a repeated subject line.
Keep the recipient set temporary and outside recurring application jobs. Stop the bounded workflow after the planned send. No message is sent by this article, and no native Resend integration with LDM is assumed.
Interpret three evidence layers
First, simulation can demonstrate handling of the predefined scenario. Second, sender logs can establish actual submission and transport events for a real message. Third, LDM can report the location where a received copy was observed at a particular time. Preserve these layers independently in the record.
A copy in Spam does not mean your event handler failed. A missing copy does not establish Spam or a complaint. A small panel also cannot represent every customer's history, rules or future interaction. Use the evidence to choose the next diagnostic check without promising universal Inbox placement.
Frequently asked questions
Does delivered@resend.dev measure Gmail Inbox placement?
No. It is a documented simulated-event recipient. A real mailbox observation requires a different, permitted real send.
Does complained@resend.dev prove filters disliked my content?
No. It exercises the predefined complaint scenario. Keep that result separate from content and folder investigations.
Should I disable event tests after adding placement checks?
The checks answer different questions. Keep useful application-event tests and add a bounded real-route observation where that evidence is needed.
Official references
Reviewed 2026-11-05; technical addresses are literal documented examples, not destinations used by this article.