Yes. WPML stores its strings in wp_icl_strings and translations in wp_icl_string_translations, both of which are queryable like any custom table. SleekAI can be configured to read counts and recent rows from those tables and expose them as variables in the system prompt, so the model answers with real backlog numbers instead of generic estimates.

By default, no. SleekAI is read-only against the WPML tables. The bot tells you what is missing and where it lives. If you want write actions, you can wire WPML's own translation hooks into a SleekAI tool with a capability check, but most teams keep the actual translation work inside the WPML String Translation editor for review.

It works with both. The base WPML plugin stores language and config data in icl_sitepress_settings and related options; the String Translation addon adds the strings tables. SleekAI can read any combination of those, so the bot can talk about both site-wide language settings and per-string status without separate setup.

Yes. The bot can match against the source string in wp_icl_strings using a fuzzy lookup that you wire as a SleekAI variable. Search for "add to cart" and the bot returns the matching entries with their gettext domain and translation status, even if the translator did not know the exact original wording.

Not unless you configure it that way. Display conditions on user role and capability keep the WPML String Translation bot scoped to translators and admins. A public chatbot you run on the same site is a separate bot under multibot with its own data sources, and it does not read the strings tables.

Yes. wp_icl_strings stores the gettext domain along with the string, so the bot can report counts per domain (theme, woocommerce, plugin-X). That is usually how editorial teams reason about backlogs, since one theme update often adds many strings under one domain at once.

It can read whatever you expose. WPML stores the auto-translate credit balance in a setting and updates it after each batch; SleekAI can read that value and tell an editor how many credits remain. Refilling and configuring engines still happens inside WPML's translation management screens for safety.

Yes. Each WordPress install has its own WPML database and its own SleekAI configuration. Staging often has fewer translated languages or a smaller string set; the staging bot can talk about that without leaking production backlog numbers. Conversation logs stay per environment, so the data sets do not mix.