A documented AI-assisted JRPG localization release burst. One production workflow moved difficult game data through extraction, translation, rebuilding, runtime QA and public patch distribution across three distinct JRPGs and four localization outputs.
The time evidence is layered, not blended. Public announcements, GitHub publication timestamps and production stages are separate clocks. Toggle them to see exactly what the headline measures.
Xenosaga appears twice because the rough-English RC1 and the later complete Dutch pass are separate localization outputs from the same Nintendo DS game.
Core/story machine-readable text reached rough English across the supported layer. Runtime QA proved boot, dialogue, opening choices and battle entry, while the serializer relocated 1,304 EVC branch targets after variable-length English text moved bytecode.
An existing incomplete public English basis was pushed through corpus closure, RMD compilation, glyph reconstruction and deterministic CPK rebuilding into a rough-complete runtime-text release. This is a completion case, not a claim that an untouched game was translated from zero.
A full story-playable Android English fan translation, with main story, opening narration, core menus, battle instructions, cards, items, equipment, character surfaces, save/load, map, Study, results and other gameplay text localized. The beginning-of-game video is runtime evidence, not the translation boundary.
Known boundary: some splash-screen artwork and scattered peripheral Korean can remain. Story-playable does not mean every-raster-complete.
The same game became a second, broader localization output: full Dutch dialogue/story, choices, menus, system text, graphical UI, main title menu and opening narration. Cold-boot tracing found the final Japanese location card in an alternate runtime source and corrected it to Kenia, Turkanameer.
A language model can emit strings quickly. A public localization still has to survive IDs, pointers, opcodes, fonts, images, containers, rebuilds, runtime failures, QA, residue scans and patch distribution.
The attraction keeps artifact identity next to the narrative. Commercial ROMs/APKs are not distributed by these release records; patches or builders require compatible originals.
The August 28 research stays frozen as a pre-release snapshot. It identified Xenosaga, Nova and Summon Night 3 as high-value targets. The next days changed their state without changing the old research retroactively.
The survey does not manufacture a record by changing definitions. Single-game rapid MTL, long AI-assisted localization programs and multi-title workflows are retained as comparators, with unknowns left unknown. The earlier-precedent cutoff is the first canonical burst announcement at 29 Aug 2026 16:25:16Z; later projects such as the Aug 30 Super Robot Wars Z release are comparison-only.
| Project | Game(s) | AI / MT method | Translation time | Total production / release time | Public patch | Confidence | Source |
|---|---|---|---|---|---|---|---|
| Super Robot Wars Z | 1 | Claude, technical + translation | ~2 days bulk | ~18 days reported total; ~16 days repair/editing | Yes · Aug 30, 2026 | STRONG EVIDENCE | OPEN ↗ |
| SRW Z2 Hakai-hen + Saisei-hen | 2 | AI bulk + creator cleanup | Unknown | Public outputs spread across months | Yes | STRONG EVIDENCE | OPEN ↗ |
| SEGAGAGA | 1 | DeepL + ChatGPT base, human translator review | Unknown | Months-long review + testing | Yes · Feb 27, 2026 | CONFIRMED | OPEN ↗ |
| Yo-kai Watch 4++ | 1 | Google Translate-backed all-string mod + scripts | Unknown | Unknown | Yes | CONFIRMED | OPEN ↗ |
| Kuro no Kiseki II | 1 | Reported DeepL Korean → English | Same-day text conversion reported | Full localization production time unknown | Reported | COMMUNITY REPORT | OPEN ↗ |
| Dragon Quest X Offline | 1 | Hybrid ported/MT/revised Unreal pipeline | Unknown | Long-running / unknown | Yes | CONFIRMED | OPEN ↗ |
| Growlanser VI | 1 | Human fan translation + LLM-assisted engineering | Unknown | Years-long project | Partial releases | CONFIRMED | OPEN ↗ |
| Langrisser V | 1 | AI-assisted full script localization | Unknown | Unknown | Yes | STRONG EVIDENCE | OPEN ↗ |
| Monster Hunter Portable 3rd | 1 | Gemini heavy-lifting + creator review | Unknown | Late-2025 restart → Jan 31, 2026 completion | Yes | STRONG EVIDENCE | OPEN ↗ |
| Grandia: Digital Museum | 1 | AI-assisted fan translation | Unknown | Alpha release, incomplete | Yes · Jul 6, 2026 | CONFIRMED | OPEN ↗ |
No earlier documented precedent was found for this specific combination of multiple traditional JRPG titles + substantive localization outputs + public patches + an approximately 77-hour public announcement window. This is a scoped survey result, not a world-first claim.
The newest verified production state has the historical STORY exact-source denominator fully translated and width/capacity-prepared. That does not mean the whole game is released: rebuild/runtime QA and remaining system, battle, tutorial and graphics work stay separate factory stages.
Every stronger interpretation stays above a lower hard-fact layer. The final rung remains deliberately unused.