OpenFront and Steel Tide: two browser strategy games with almost nothing in common
Both open in a tab, both are free, both are real-time strategy. After that they disagree about everything — what you control, how many people are in the match, and what winning even means. An honest comparison.
If you searched for a strategy game you can play in a browser this year, you found OpenFront before you found us. It is the most-played real-time strategy game on the web by a very wide margin, it is open source, and it is about to arrive on Steam. It is also, mechanically, about as far from Steel Tide as two games in the same genre can be.
That makes it a useful comparison. Not "which is better" — they are not competing for the same twenty minutes — but "which twenty minutes do you actually want".
What OpenFront is
OpenFront.io is a real-time territory game. You spawn as a dot on a map drawn from real geography, and you expand by sending troops into the land next to you. Two sliders run your whole nation: one splits your population between workers who earn gold and troops who fight, the other decides what fraction of your troops each attack commits. Borders move; there are no individual units to click on.
Gold buys buildings — cities that raise your population, ports that trade, factories, defence posts, SAM launchers, missile silos — and eventually it buys the thing the game is famous for. An atom bomb costs 750,000 gold, a hydrogen bomb five million, and a MIRV considerably more. Late games are decided by who has silos and who has SAMs.
The other half of the game is people. A public lobby seats up to a hundred humans, padded out with a few hundred bots, and the map is far too big to win alone. So you ally, and alliances expire after about five minutes unless renewed, and breaking one halves your defence for a while afterwards, which is exactly long enough for your former friend to make a point. You win a free-for-all by holding 80% of the land; a team needs 95%. A match runs twenty to fifty minutes.
The project is a fork of WarFront.io, in the lineage of Territorial.io, written in TypeScript and published under the AGPL with its assets under Creative Commons. The team's own announcement puts it at around 200,000 players a day, and the Steam Early Access version, with cross-play against the browser build, is dated 17 September 2026. Mobile apps exist as well.
None of that is faint praise. A hundred-player real-time game with diplomacy that runs in a tab is a genuine design achievement, and it is popular for the plain reason that it is good.
What Steel Tide is
Steel Tide is the other kind of RTS: the one with a base. You start with a headquarters and a couple of engineers, place extractors on metal deposits, keep the power on, build a war factory, and produce tanks that you personally point at things. Thirty-two units across land, sea and air, twenty-eight buildings, and one counter table that says what each weapon does to each armor class.
A match holds up to six seats in any mix of humans and AI, in teams or free-for-all, on fourteen authored maps or a generated one. A skirmish resolves in fifteen to thirty minutes. You are out when you can no longer rebuild — no headquarters, no production building, no engineer — and an army on the field buys you ninety seconds to fix that.
It also opens in a tab, with no download and no account, and the same bundle ships as a desktop app and as phone apps. Multiplayer runs on an authoritative server that you can install with one command.
Side by side
| OpenFront | Steel Tide | |
|---|---|---|
| what you control | a border and two sliders | individual units and buildings |
| players in a match | up to ~100, plus bots | up to 6, human or AI |
| economy | gold; workers versus troops | metal from extractors; power from plants |
| combat | attack ratio against a neighbour's troops | 32 unit types, weapon class versus armor class |
| diplomacy | alliances that expire, betrayal with a penalty | teams fixed at setup, chat |
| the big button | atom bomb, hydrogen bomb, MIRV | a nuclear silo and the interrupter that shoots warheads down |
| match length | 20–50 minutes | 15–30 minutes |
| win condition | 80% of the land | every rival unable to rebuild |
| solo play | practice against bots | six-mission campaign, three AI grades |
| install | none; Steam Early Access from 17 September 2026; mobile apps | none; installs as a web app; desktop and phone apps |
| hosting | open source, run your own | one-line dedicated server, LAN from the app |
Where each one wins
OpenFront wins on people. A hundred-player free-for-all produces stories a six-player game cannot: the coalition that forms against whoever is winning, the neighbour who was an ally for exactly five minutes. Diplomacy is the game, and everything else exists to give diplomacy stakes.
OpenFront wins on onboarding. Two sliders and a click on a border is a complete tutorial. You are playing in ten seconds, and you understand what you are doing in thirty.
Steel Tide wins on the army. If what you enjoy about strategy games is the unit — choosing it, positioning it, watching a tank destroyer delete the heavy armour that was about to roll your line — then a game where combat is a number moving between two territories is not going to give you that. Steel Tide is about the composition of a force and where you put it.
Steel Tide wins on the base. There is an economy to grow, a power grid to keep out of deficit, a tech tree to climb and a production queue to keep full. The five habits that make an RTS player good are all present because the game has the systems those habits act on.
Steel Tide wins on playing alone. Bots in OpenFront are filler between humans. Steel Tide's AI opponents scout, expand, counter what they have seen of your army and attack in escalating waves, and there is a campaign. A match with no other human in it is a real match.
What we would take from OpenFront
Two things, honestly.
The first is the lobby. OpenFront's public games fill continuously and nobody is ever waiting for a friend to come online. Steel Tide's multiplayer is a server you or someone you know runs, which is the right model for a game about six people and the wrong model for a game about a hundred — but it does mean the moment of "who do I play with?" is ours to solve, and we have not solved it yet.
The second is alliances that expire. A time-limited pact that has to be actively renewed is a lovely piece of design: it makes trust a decision you take repeatedly rather than once. Steel Tide's teams are fixed when the match is set up. That is correct for a game this size, and it is still a little less interesting.
Which twenty minutes
If you want a crowd, diplomacy, and the particular joy of nuking someone who deserved it, play OpenFront. It is excellent and it costs nothing.
If you want a base, an army and an opponent whose composition you have to answer, that is the game we make. It also costs nothing, it also opens in a tab, and the first ten minutes are written down.