Customers translating a large help center (49 articles) with an AI assistant through the Gleap MCP have to split the work across many conversations because the help center tools return very large payloads. Current pain points: 1. get_help_center_article returns the full editor JSON (IDs and attributes on every node) for every language, plus a plain-text copy. 2. update_help_center_article requires the full article as editor JSON for the language being updated. No Markdown input and no partial/section edit. 3. Every successful update echoes the full article in all languages plus plain-text copies. 4. get_help_center_articles does not show which languages each article has, so articles missing a translation are hard to find. Requested: - Language filter on get_help_center_article (e.g. return only "en"). - Markdown option for reading and writing article content. - Compact update response (success, version, updated languages, no content echo). - Partial edit option, e.g. find-and-replace within one language of an article. - Available languages per article in get_help_center_articles, or a filter such as "missing language: fr". Underlying need: translate a whole help center in one AI conversation instead of six or seven, with less risk of long articles failing partway.