Where things stand
If you've been following along (hi, both of you), the backend is done.
All seven microservices — Identity, Gamestate, Catalog, Party, Instance, Chat, and Admin — are implemented and passing a full 4-player smoke test. The smoke test covers the whole flow: register four accounts, create four characters, form a party, run a ready check, request a dungeon server, verify join tickets, open loot chests, and load character data on the simulated server side. Twelve steps, all green.
This is the part of the game players will never think about. It's also the part that would've caused problems in everything downstream if I got it wrong, so I'm glad it's solid before touching Unreal.
The dungeon design is locked
Wild Grotto — the first dungeon, designed for level 15 parties — has a complete design doc. Four zones, eight enemy types, four boss fights, a crafting mechanic inside the dungeon, and zone gates that open when the relevant boss dies.
The bosses:
- Kritisk, the Razor-Leaf — a mantis with a counter window. Attack during Silver Stance and you get marked and jumped. The mechanic punishes not paying attention.
- Guttermaw — a toad that stacks toxin on the whole party over the course of the fight. There's a crafting station nearby. If you explore before pulling, you can make antidotes. If you don't, you'll find out what happens.
- Fang & Claw, the Packlords — a wolf and a bear fought at the same time. They share a Blood Bond: when one dies, the other enrages. There's a correct kill order. It's not obvious.
- Vylos, Nature Elemental — the final boss. Stationary, ranged, introduces a new mechanic at 75%, 50%, and 33% HP. The last phase goes badly if the party hasn't figured out Symbiotic Link by then.
The design goal for all of them is that a wipe teaches you something specific — not just "try again and maybe survive."
What's left
Client work. Models, animations, blueprints, sound, level design. That list is long and I'm not pretending otherwise. But having the backend done and the dungeon fully specced before opening Unreal means every decision I make there is implementing something that already exists on paper, not designing and building at the same time. That's a better position to be in than I expected to be at this point.
Playtest registration opens when a full party can run Wild Grotto start-to-finish. I'll have a better read on the timeline next month.
On building this alone
Someone asked recently if solo development is sustainable for a project this size. Honest answer: I don't know yet. What I do know is that the approach of fully speccing systems before building them has worked so far. We'll see how it holds up in Unreal.
More next month.
— Kevin