Rules

The automatic rules you create to decide what to do with each comment: reply, hide, delete, ban, invite to DM.

Help Center / Rules
How do GeniReply's rules work?

GeniReply starts with an empty Rules list: you build them based on your business and community needs. Each rule you create has:

  • Number and Name — to identify it (you choose the name based on the use case it covers)
  • Priority — from 0 to 100. When matches tie, the highest priority wins
  • Keywords — terms that trigger the rule (e.g. "how much does it cost", "scam")
  • Exceptions — words/emojis that flip the behavior (e.g. 🤡🤮 triggers the "exception actions" instead of the standard ones)
  • Standard actions and Exception actions — see the dedicated question

The AI Haiku classifier combines keywords with semantic understanding: it can match even when the user uses synonyms or non-exact phrasing.

What actions can I configure on a rule?

On each rule you can enable one or more of these 5 actions:

  • 💬 Reply publicly — publishes a reply comment under the original one
  • 🙈 Hide the comment — temporary soft-hide, useful when GeniReply is uncertain and wants to ask for guidance
  • 🗑️ Delete the comment — permanent removal from the feed (offensive content, spam)
  • 🚫 Ban the user — blocks the author on the page
  • ✉️ Invite to DM (Direct Message) — sends a private message with custom text

Actions can be combined: for example, a rule that catches phone numbers left in comments can hide the comment (for privacy) and at the same time send a DM to the user inviting them to write privately.

How do I create a new rule?

Two paths:

  1. From the Rules menu → + New rule button. Fill in number, name, keywords, exceptions, actions, reply text and DM text.
  2. From a real comment in the Inbox → Create rule from this comment button. The comment text is used as the base and you can refine it.

Before saving, test the rule in the Simulator with 2–3 variants of the message you want to intercept. If the confidence is low, enrich the keywords.

⚠️ Avoid duplicate priorities: if two rules share the same priority and match the same text, the winning order is not deterministic. GeniReply warns you in real time: when creating or editing a rule, if the priority you pick is already used by another rule, a yellow message appears under the Priority field — "⚠️ Priority X is already assigned to: 'Rule name'". You can still save (it's just a warning), but we recommend picking a different value so behavior stays predictable.
How does priority between rules work?

Priority is a number between 0 and 100. When a comment matches several rules at the same time, GeniReply applies only the one with the highest priority.

Example: if you have a specific rule with priority 100 and a more generic rule with priority 40, and a comment triggers both, only the specific one (100) fires. This mechanism lets the most specific rules (100) always "prevail" over generic ones (40).

💡 When you create a new rule, set priority high (80–100) for narrow rules with very selective keywords, and low (10–40) for catch-all rules (e.g. default emoji replies) that should only fire if nothing more specific matched.
What are 'exceptions' and when do I use them?

Exceptions are words/emojis that flip a rule's action. Typical case: the "Emoji" rule replies warmly to 👍❤️ but if the user writes 🤡🤮 you need to delete+ban.

How to configure them:

  1. Open the rule in edit mode
  2. In the Exceptions section, enter the words or emojis that should trigger the alternative behavior
  3. In Exception actions choose what to do (e.g. 🗑️ Delete + 🚫 Ban)

When the AI classifies a comment, if the text contains a word listed in the exceptions it applies the exception actions; otherwise it applies the standard ones.