Presets

A preset is a finished starting point for one kind of content. Choose one when you create a template and its fields, policies, triggers and content rules are filled in, most with AI instructions too. You adjust what does not fit and publish.

From then on the template is yours. Nothing ties it to the preset, and a later change to a preset does not reach templates already made from it.

Open Templates, choose New, name the template and pick the preset under Start from. The slug comes from the name you type, not from the preset.

The eight presets

Blank template is the ninth choice: no fields and no policies.

A narrower kind of content is a preset with your own words. A used-car listing is the Listing preset with a different What this content is, other field descriptions and other AI instructions. Every preset fills in What this content is, and a description for nearly every field.

Fields of each preset

A max size is reported and does not reject, as Fields and validation explains.

What each policy does

In every preset, toxicity, hate_speech, profanity, violence, self_harm, sexual and illicit flag, and lookalike_characters sends to review. The other policies differ from preset to preset:

Flag and review are the effects of a trigger. An observed policy is scored and returned: it fires no trigger and adds nothing to severity. Presets that carry AI instructions flag on them from 0.70.

Some policies fire on one field only:

• Listing: links_contact in title, profanity in description, and made_with_ai on photos, from 0.90.
• Job posting: links_contact in title.
• Profile: profanity and lookalike_characters in display_name, links_contact in bio.
• Forum thread: profanity in title.

A policy with a field trigger does not fire elsewhere, so a link in a Listing description does not flag. It still counts for severity, and the listing can be sent to review or rejected all the same. Triggers explains why and how to cover every field.

Four presets also set options:

• Listing and Job posting count steering off the platform, such as "write me on WhatsApp", as a contact.
• duplicate looks back 7 days in Listing, at texts of 40 characters or more, and in Forum thread, at texts of 80 or more. In Job posting it looks back 14 days, at texts of 60 or more.
• In Message, duplicate looks back one day, at texts of 60 characters or more by the same author.

Content rules of each preset

Content rules are read in order and the first one that matches decides. Where none matches, severity triage does, with the default thresholds: review from 0.40, reject from 0.80.

Read the Listing rules with trust levels in mind. A link from an author at trust level new is sent to review and one from regular up is allowed. The level in between, basic, matches neither rule, and triage rejects the link.

AI instructions of each preset

The full text is on the Instructions tab of a template made from the preset.

Fitted to your plan

Five presets carry it: Listing, Review, Forum thread, Comment and Social post. After an upgrade it is not added back by itself. Switch it on in the Policies tab of the template.

Next

• TemplatesA template is the configuration for one kind of content. It holds the fields, the policies and the rules, and requests name it by its slug.
• Fields and validationA template's fields say what a submission must contain. Validation runs first, costs nothing and answers with one error per problem, named by field.
• TriggersA trigger says what a policy's score sets off, a flag or a review, for the whole template or for one field. One of your own replaces the default.
• Content rulesOrdered rules on a template turn scores, author trust and field values into allow, review or reject. Severity triage decides when no rule matches.