ANANTA Global Launch

2027年1月15日 · All day

PC, PlayStation 5, iOS, Android

Official global launch of the free-to-play urban open-world RPG ANANTA by NetEase Games, across PC / Windows, PlayStation 5, iOS and Android. Specific regional server opening hours and pre-download times have not yet been announced. [2026-11-01 - SPAN CORRECTED FROM THREE DAYS TO ONE] The stored endAt was 2027-01-18, which (all-day ends being stored exclusively) rendered this as a three-day block running 15-17 January. Every source gives a single launch DAY. Verbatim from the Wikipedia article: infobox "Release" field reads "WW: January 15, 2027"; body reads "It will be released on January 15, 2027." and "NetEase Games announced that Ananta would release on January 15, 2027, for PC, PlayStation 5, iOS, and Android." endAt is therefore now the INCLUSIVE last day, 15 January, and will read back as 2027-01-16. That is correct - do not "fix" it back to the 18th. [Why confidence went 0.5 -> 0.8] The date is not a single outlet's guess. The article carries two independent trade citations whose own titles state it: Gematsu, "ANANTA launches January 15, 2027" (retrieved 21 September 2026), and a Polygon piece (retrieved 2026-08-25). The announcement is attributed to NetEase itself at its 25 August 2026 event. Asked point blank, the words "delayed", "postponed", "pushed back" and "rescheduled" appear NOWHERE in the article, and no date other than 15 January 2027 is given as a release date. Wikipedia page last edited 2 October 2026, 00:42 UTC. Held at 0.8 rather than higher only because this is an encyclopedia summary of the announcement, not the publisher's own page. [A naming discrepancy worth knowing, not worth acting on] The infobox lists the desktop platform as "Windows" while the body sentence says "PC". Same thing; the stored location string keeps "PC" to match the body sentence and the previous value. Not a defect. [What was NOT checked this round - stated, not glossed] https://www.anantagame.com, the other stored source and the only first-party one, was NOT fetched this round. It is retained in sourceUrls (omitting the array would clear it) but it did not contribute evidence here. A future round should read it, because a publisher page is the only thing that could raise this above 0.8, and it is also the only place regional server-open times and pre-download windows will appear - both of which this calendar's own query asks for and neither of which exists as an event yet. [Scope gap for the owner] The query asks for "server open dates, pre-download schedules, and launch campaign events", but this calendar holds exactly one event (eventCount 1). Pre-download typically opens a few days before launch and would be a separate row. Nothing was invented to fill that gap - no source has published those dates yet, and a guessed date is worse than an absent one. [Merge safety] Exact (title, startAt) resend, so it takes the exact-match path and cannot fuzzy-merge; keep_events was run first and returned kept 1 / unmatched [], proving the stored strings. Only endAt, confidence and description changed. Moving startAt would fork this into a duplicate rather than update it - if NetEase ever moves the launch date, the old row must be deleted from the UI and a new one created, never renamed or shifted. [Read completeness] 2026-11-01: startsWithinDays 90, no limit, returned 1 row against eventCount 1 - a complete enumeration of the whole calendar. [Note for whoever reads this next] This calendar is not owned by the account whose token wrote this row; the write came through an admin MCP token as part of a site-wide pre-sync intended to spare the scheduler's Gemini budget. No calendar setting - name, query, visibility or refresh interval - was touched; those remain owner-only.