The core function module provided by AnQi CMS, helping you quickly build an enterprise website
Functional Icon
Email reminders should be configured by event grouping: the publication of new articles belongs to write events, triggered by the content status turning into a formal document; receiving comments belongs to review events, triggered by comments entering the content review queue; backup and deployment actions belong to exception events. Actions such as draft saving and moving to the recycle bin should not trigger external reminders, and recipients should be grouped by responsibility and merged for noise reduction.
Functional Icon
Should a separate content model be built for the marketing activity page, judged by the number of times the structure is reused: One-time activities can be carried by a single page, with the navigation entrance being enough; for serialized activities with fixed fields, a custom content model and custom fields should be built. This article explains the granularity of modeling, the division of labor between single pages and columns and navigation, the use of scheduled publishing to control the online rhythm, and the three signals for splitting out the second set of models.
Functional Icon
The built-in AI skill covers content planning, search analysis, batch operations, site health checks and troubleshooting. It can provide conclusions such as address and jump, site map and submission, and whether the backup is the latest. This article explains what the skill can find, why the conclusions found need to be manually reviewed, and how to proceed with the confirmation process when involving writing operations, and provides a manual review checklist.
Functional Icon
Separating common questions from articles and building a separate content model for them brings benefits mainly from three points: a self-contained Q&A, unified field structure, and stable address format. These three points make Q&A search easier to hit; whether the model recognizes the name 'common questions' does not depend on how the fields are built. This article explains the choice between building a separate model and placing it in an article, as well as how the keyword library and site description file work together in such content.
Functional Icon
When launching the first version, it is recommended to open the background modules related to search and the large model in three stages: pseudo-static addresses, site maps, and active push belong to the first batch of must-open items, followed by the site description documents for the large model; 301 redirects and exclusion files are determined according to whether the version is revised; anti-crawling interference codes, scheduled publishing, and anchor text can be postponed. This article lists the functions and switch order of these modules in stages, and explains which conclusions still need to be reviewed manually
Functional Icon
The default location specified in the proposal is under the site root path at llms.txt, and it is also allowed to be placed in any sub-path; the necessary part of the file is only the title of the site name, the rest are optional content organization. In the site cluster scenario, a separate one should be placed for each site: the names, summaries, and content lists are all different, and sharing one will direct the read model to the address of another site. The site supports automatic generation of this description file, and after changes, it is updated by the list synchronization mechanism.
Functional Icon
Can be separated, and it is separated by default. Under multi-site management, keyword libraries, anchor text entries, and full-site replacement are read and written according to the current site, each site has its own words and link targets, and the words of Site A will not appear in the automatic links of Site B. The two places that truly cause word confusion are: when the same word has different meanings under different brands and is still matched for placement according to the literal meaning, and when the full-site replacement selects the wrong scope of objects. The way to avoid this is to build libraries separately by site, limit the placement objects for anchor text, and estimate the impact before replacement.
Functional Icon
It can be called continuously, but the modification actions are not exempt from confirmation. The AI tool set in the station is organized by the intent directory, reading actions can be called continuously by the intelligent agent, and writing actions are merged into one confirmation and executed in the same round, so multiple-step tasks will not pop up a window for each step. The approval scope is set according to the intent domain, the default high-risk domain at the site level is not enabled, and it needs to be explicitly opened. The access party is the built-in AI Chat and external AI client.
Functional Icon
To check if the write path is the same. The batch import of ZIP archives and Excel tables, as well as the import/archive import interface, all follow the same saving process as the background editor: the required fields such as title and category are still checked, the address identifier is still generated by removing duplicates from the existing content, and there is also an independent authentication pre-check on the interface side. What really needs to be handled separately is the duplicate check and override strategy, as well as the scope of operation of the input and output layers, rather than copying the check and attaching it to the import.
Functional Icon
They can be recovered. By default, deleting an official document changes the status instead of erasing the data: it carries the original text, title, category, tags, and address identifier into the recycling bin. Restoring is just changing the status back, no need to edit it again. Items in the recycling bin that exceed the retention period will be cleaned up regularly. Physical deletion cannot be saved in such cases, and for batch accidental deletions or site-level losses, the backup and recovery path must be taken. The difference is in the scope of recovery, not the difficulty of recovery.
Functional Icon
The scheduled release and the pending release are not the same thing, nor are they the fifth and sixth states in parallel: the pending release is one of the four states of the document, and the scheduled release is responsible for handing over the step of 'pending release to formal' to the time point for automatic triggering. Set a content time later than the current time and save the content, it will enter the pending release, and be cut into a formal document at the appointed time; it is not visible to the outside world before it is triggered. Drafts can be viewed with preview parameters, and deleted content can still be restored in the recycle bin.
Functional Icon
There are three common forms of address arrangement for multilingual websites: path prefix under the same domain, subdomain split by language, and independent domain for each language. The path prefix has the least maintenance, and it is the easiest to express the switching between languages; subdomain and independent domain will split the host ownership, and configuration and verification need to be done separately. The site supports configuration and switching of multilingual sites, full-page HTML translation, and the pseudo-static rules should be able to cover all language paths.
