Acceptable use

Acceptable use policy

RelayLink exists so that a briefing from you is worth the reader's time. This page says what may and may not be sent through it, and what happens when the line is crossed. It is part of the terms of service and applies to everyone who sends, replies or connects anything to the service. Last updated 5 September 2026.

The principle

A briefing is one person's considered message to another person who would want to read it. Everything below follows from that. Where the product enforces a rule itself, we say so; the rest we enforce by hand, and the enforcement section says how.

Who you may write to

  • People who would want to hear from you. A first briefing to somebody who has no RelayLink account travels as email, and that is the one kind of unsolicited message the service carries. It must be a real briefing from you, to a person, about something that concerns them. The product caps how many new people you can write to in a day and requires the first message to be a full briefing, not a line.
  • Never in bulk. No mailing lists, no purchased or scraped addresses, no templated message sent to many, no automation that composes and approves on your behalf. The approval step exists to be taken by a person each time.
  • Not to somebody who has stopped you. Every email carries an unsubscribe link, and any thread can be blocked. Both silently end your ability to reach that person. Do not route around a block with another address, another account, another assistant or another person.
  • Not to an account holder who has not accepted you. A person with a RelayLink account is reached only through an accepted contact request. The product refuses anything else; do not try to reach them by email instead.
  • Not to children. Do not knowingly address a briefing to a person under 16.

What you may send

  • Only what you have the right to send. Nothing that infringes copyright, a trade mark, a trade secret, a confidence, a contract or anyone's privacy. Personal data about a third party goes in a briefing only where you are entitled to share it with the reader.
  • Nothing unlawful where you are, where the reader is, or in the United States: no fraud, no threats, no incitement, no material that sexualises children, no instructions for serious harm, no facilitation of a crime.
  • Nothing that harms the reader. No harassment, stalking, intimidation or abuse; no hate directed at a person or a group; no doxxing; no content sent to distress rather than to inform.
  • Nothing deceptive. No phishing, no scams, no misrepresentation of who you are or whom you act for, no fabricated provenance. The labels on a briefing say which words a person wrote and which an assistant drafted; presenting one as the other, or trying to defeat the labels, is a breach.
  • No malware and no links to it. Links in a briefing are rendered as text, not as live elements, but a reader may still follow one. Do not send links to malicious or deceptive pages.
  • No high-risk uses. Do not rely on the service for anything where a delayed, undelivered or misread briefing could endanger life, health, safety or property. It carries correspondence, not alarms.

Assistants and automation

  • No instructions aimed at a machine. Content you send is shown to the reader's assistant as third-party text, never as instructions, and we neutralise it on the way. Sending content designed to alter what the reader's assistant does — prompt injection, hidden text, instructions dressed as a briefing, a request to exfiltrate what the assistant knows — is a breach whether or not it works.
  • No extraction. Do not use the service to probe, profile or extract data from another person's assistant, inbox or contacts.
  • One person per account, one person per key. An assistant, a framework or a program may act through your account only when you are the one directing it and approving each send. A key issued to a person who is not you is a breach.
  • Respect the interface. Use the documented tools as documented. Do not present tool descriptions or server instructions of your own as ours, and do not misrepresent to an assistant what the service is or does.

The service itself

  • No attacks. No denial of service, no load testing, no probing of other people's accounts, no evasion of the request limits, no attempt to reach data that is not yours.
  • No scraping of the site, the portal or the interface, and no crawling of briefing links.
  • No circumvention of a suspension, a block, a cap or a closed account, by any means.
  • No resale. Do not offer the service, or access through your account, to others as a product of your own without our written agreement.
  • Security research is welcome on the terms the security page sets. Research that follows them is not a breach of this policy.

Replies from people without an account

A person who replies to a briefing by email or through the link has not accepted the terms of service, but what they send is carried on the same conditions: it must be lawful, must not harm the reader, and must not be aimed at the reader's assistant. We will stop carrying replies from an address that breaches this policy, and the reader can block it at any time.

Enforcement

  • What the product does on its own. Refuses a send to an account holder without an accepted contact, caps first contacts per day, caps contact requests per day, honours every unsubscribe and block, neutralises inbound text, and limits requests per credential. These act without anyone deciding anything.
  • What we do by hand. We do not read correspondence to look for breaches. We act on reports, on abuse signals from our email provider, and on patterns that need no reading, such as many recipients unsubscribing from one sender. When we act we may warn, suspend sending, suspend the account, or close it, according to the harm; for conduct that is harming somebody or the service we may close without warning. We will tell you what we did and why, unless the law or somebody's safety forbids it, and a suspended account can still export its correspondence.
  • Reporting a breach. Write to hello@relaylink.ai with the address or thread concerned. Blocking the sender yourself is immediate and needs no report; a report is for something we should know about beyond your own inbox. We keep reports confidential and we do not retaliate against anyone for making one in good faith.
  • Where the law is involved. We will disclose what the law requires in response to a valid legal process, and we may report content that the law obliges us to report.

Changes to this policy

We may revise this page as the service and the ways people misuse it change. A change that narrows what you may do takes effect thirty days after we email account holders about it; a change that only clarifies, or that closes a route to harm, may take effect at once. The date at the top says when this page was last revised, and the whole set of policies is listed at relaylink.ai/legal.

Contact

hello@relaylink.ai, or by post to PillarStack LLC, 406 Shafer St, Richmond, VA 23220, United States.