Vaelthi Elegia and the playable prologue
Neri grows up in Calvessa, a canal town in a country where rain is part of ordinary life. The opening of Vaelthi Elegia follows four days there, before a raid ends that life. Neri is turned to stone and wakes a thousand years later in a country where most of the rain has vanished.
The game is being built in Godot. There is already a playable prologue, currently estimated at roughly one to two hours. It has the major story events and gameplay systems in place, although the presentation and dialogue are still being revised.
What you can play in the current build
Calvessa has houses, canals and people following routines. You can explore it, speak with its residents and move through events that change with the day.
Combat is turn-based. Landing a finishing blow lets Neri take an enemy’s echo. Some echoes give you combat abilities; others let you interact with the world, including opening blocked routes and finding things that would otherwise stay hidden.
The story and every character’s dialogue are written, and the game plays from start to finish as authored. There is also an optional language-model component, experimental at this stage. I intend it to be a core part of the game and more than questions and answers: a way to build relationships with characters, swap stories, and feed into quest decisions and the ending. It stays optional. With it switched off, the game uses written fallbacks that always play the same way, and in either case the game’s rules decide what happens when you recruit someone or complete a quest.
The recruitment and community systems extend beyond the prologue. People can join you when their conditions are met, and the Caravanserai gives recruits a place to live and work. Staffed facilities use the skills of the people assigned to them, so building a party also changes what your settlement can do.
Working on the visual style
An individual painting can look convincing while being of little use to a game. The same character needs to work from several directions, in different lighting, alongside props and buildings made through different processes. A good still doesn’t answer what happens when the camera moves.
The visual target is an illustrated, painterly world with 3D environments and characters that belong in them. I use generative image models to make concept paintings and asset candidates, then compare them with defined art references and revise or reject the results. The artwork on this site includes development illustrations, which are labelled separately from gameplay captures.
I’m also developing Vaelthi Tools to connect parts of the asset process, including image editing, object extraction and 3D preparation, with review and the game. Individual components exist, but the complete workflow still needs to be tested in use.
The next updates will show Calvessa in Godot beside the painted references it is aiming for, and how the characters hold up once they are moving through a scene.