Hi everyone.
It has been a while since we released the Highland Keep teaser. The response went far beyond what we expected, and Highland Keep has now passed 100,000 wishlists. Thank you! In our previous devlog, I mentioned that we were looking for a way to make development sustainable in the long term. We have now secured funding that allows us to work on Highland Keep full-time. Our team has also grown from two people to four, with a programmer and a 3D artist joining us. There is still a lot of work ahead of us, but I wanted to show you what we have mainly been building over the last few months. Most of the work below supports one goal: creating a settlement that continues to live and function, whether you are managing it from above, exploring the world yourself, or preparing for an attack.
Highland Keep takes place on one continuous map. Every human and animal remains part of the simulation, even when off-screen.
We are designing the game around roughly 5,000 units, with distant characters updating less frequently to maintain performance.
To make this possible, we built BurstCanvas, our custom visual AI system. It defines how characters move between work, rest, combat, player orders and other behaviors.
Our previous solution became too expensive at scale. BurstCanvas is lighter and can process many characters in parallel, giving us the foundation needed to reach our target population.
We’re generally a bit obsessed with performance. Personally, I’d really like HighlandKeep to run smoothly on a card like the GTX 1080, with quality close to what we’re showing in the screenshots and trailer.
Also, I can confirm that Highland Keep will be available on Linux.
Our world population system keeps track of the target population for each species and looks for sensible places where groups can live or recover over time. It considers terrain and habitat, whether an area is dry, wet or near a shore, whether it is reachable, how close it is to other animals or people.
The behaviour of the animals themselves is built from the same foundations as the rest of our AI. Grazing animals take short steps through suitable terrain and remain connected to their group. They react to threats and try to find safe routes rather than only running in a straight line.
Wolves use more aggressive group behaviour. They roam around their pack, maintain some cohesion, react to nearby threats and can circle or lunge when attacking instead of every wolf blindly charging at the same point.
We also end up having custom NavMesh.
Trees, buildings and barricades can all change the available routes. Attackers can compare a long path around with breaking through a door or barricade, allowing them to look for better ways into your castle instead of blindly attacking the nearest wall.
Most deep water has a strong current that pushes the player around upon entry. Being in the water also rapidly depletes stamina, so the player eventually begins to drown.
Underwater swimming is not supported.
Water system prepares gameplay data describing things such as the water surface, depth, flow and shoreline. Navigation, physics and other systems can query that data directly
There is more gameplay work to do around rivers and crossings, but this gives us the foundation for water to influence movement and castle defense instead of being only a part of the scenery.
Some of you asked about tartans after the teaser. At the time, we already had a very early prototype. That system has now grown into a proper part of creating your clan.
You can choose your clan name and design its tartan using primary, secondary and accent colours. You can make a pattern of your own from scratch, or start making the tartan from one of the pattern presets available in the game.
The design is saved with your clan and can appear on kilts and other clothing parts.
As a survival base grows, it often turns into a room filled with identical storage chests.
You may own a tons of wood, stone and food, but none of that wealth is visible in the settlement.
In the early versions of HK, we noticed that it looked lifeless, making it difficult to create the impression of a living castle.
Currently, most "key" raw materials are stored on kinda "pallets" where their type and quantity are clearly visible.
An empty pallet can accept a stockable resource. Once the first item is stored, it remains assigned to that resource type until it is emptied.
The visible pile changes with the amount stored, so wood, stone, food and other goods can become readable parts of the settlement. Workers need to gather, carry and deposit those resources, while workstations need access to the things required for production. This makes the layout of storage and workplaces visible.
We recorded audio for shouts and various reactions a while back. Unfortunately, the workload is so heavy that we’ll likely use an external studio.
We have rebuilt how Highland Keep handles spatial audio, including how walls, roofs and enclosed spaces affect the sounds reaching the player.
Our main focus now is bringing unit management and castle defence into one complete, readable flow.
There is a reason relatively few games combine third-person survival, settlement management, large-scale simulation and tactical castle defence. We are experimenting with dedicated "defense design" mode to allow players precisly arrange units in "alarm" mode.
Every part touches several others. That makes development slower at times, but it is also exactly why we are making Highland Keep.
By the way, if you post about Highland Keep, there is a good chance we will see it on the screen in our office. Just saying :) If you have any questions, feel free to ask us on Discord. I will try to answer as many as I can. Thanks for sticking with us.
Kacper