
Quick note: The current artwork, trailer and other store-page materials are temporary and will be updated before Steam Next Fest. I wanted to put the page together quickly, so yes — AI was used for quite a few placeholder materials. But the actual ideas behind Team Frontress will be shown in the Steam Next Fest demo. So for now, please treat this page as an early work in progress. The real showcase is still coming.
After the first posts about Team Frontress, I saw a completely fair question:
What does this project actually change besides the way you join a match?
There were also several comments about the early presentation of the project — AI-generated artwork, overly broad descriptions, and the lack of proper footage from the actual build.
I partly agree with that criticism.
I started talking too early about what Team Frontress is supposed to become without showing enough of what is actually working right now. That made the project look much more vague than it really is.
So I want to clearly separate two things:
what already exists in the current alpha build;
what belongs to the core game concept and is still being developed.
To be completely transparent, I rushed quite a lot of things while trying to get Team Frontress ready for Steam Next Fest. I used AI to quickly fill some gaps, threw together parts of the presentation faster than I probably should have, and as a result some things are still rough or simply not explained very well yet.
But... it looks like I'm actually going to make it in time. ːsteamhappyː
So before Steam Next Fest, I will be cleaning up the store page, replacing the temporary artwork, improving the graphics and making the project much easier to understand from the outside.
For the festival itself, I am preparing a single-player demonstration of the main Team Frontress concept. It will not be the full persistent multiplayer war yet, but it should finally let me show the idea in a much more direct way instead of trying to explain everything through walls of text.
The current alpha is still mainly about testing the technical multiplayer foundation. The Steam Next Fest demo will be about showing what all of this is actually being built for.
The current development stage is focused primarily on Team Frontress' own multiplayer infrastructure.
Valve's original Game Coordinator is not available as open source, so Team Frontress is gradually replacing the parts of that system we need with our own implementation.
The current flow looks roughly like this:
Team Frontress client
↓
Frontress Gateway
↓
tf2pickup-frontress
↓
serveme-frontress
↓
dedicated server
This is not just a concept — the code is already present in the public repository.
Github repo
The current system handles:
Steam authentication;
matchmaking queues;
parties and team formation;
match creation;
map selection;
game server allocation;
per-match server setup;
player connection;
backfilling players into matches already in progress;
match results;
match history and basic progression.
The backend uses forks and integrations of tf2pickup and serveme.tf.
The Team Frontress gateway handles player search and match policy. tf2pickup owns durable match/player state, Source logs, RCON and match results. serveme-frontress handles game-server allocation and startup.
From the perspective of the actual shooting gameplay — mostly, yes. And that is intentional. I do not want to change weapons, classes, maps, networking, matchmaking and the global war system all at once, then try to figure out which of five different layers broke.
The first public development stage is testing the foundation:
Can an independent Team Frontress client find players, create a proper match, allocate a server, place players on teams and complete a match without relying on Valve's Game Coordinator?
So if you launch the current alpha and see familiar TF2 gameplay, that is not some hidden "final version" of the project.
It is simply the layer that is being tested first.
The main game system exists above individual matches. Instead of a match ending and disappearing completely when the scoreboard closes, its result is meant to change one shared RED vs BLU war - like in other games which uses the same mechanism with global map. (Can I mention them here? I'm new at laws)
For example, an operation may move through several stages:
BREAKTHROUGH
↓
ADVANCE
↓
ASSAULT
↓
REGION CAPTURED
Those stages can reuse existing game modes:
5CP — breakthrough;
Payload — advance;
Attack/Defend — final assault.
So player choose its side and drop to server. An attacker victory moves the operation forward. A defender victory can push it back or collapse the offensive entirely. Capturing a territory moves the frontline and opens the next part of the war.
I plan to consist players' side as much as I can. (I will modify gameplay and maps)
Not yet in the public alpha in the form I want to ship it. The repository already contains early groundwork for the strategic layer: nodes, fronts, offensive stages, population-driven front counts and a persistent event log. But in the current matchmaking branch this system is still a later development stage and is disabled by default.
I do not want to present an unfinished prototype as if it were already the finished core mechanic.