How many classes are in the main game?
The official main-game description gives the 18-class figure. The official material establishes the documented feature, while the dated gameplay examples help show how that feature appeared in a particular build. Keeping those two evidence types together makes it easier to separate a confirmed mechanic from an assumption based on another game. That statement is the strongest collected source for the advertised Early Access total.
A difference between those versions can explain apparently conflicting instructions without proving that one source is universally correct. If a step or value is not demonstrated in the collected sources, treat it as unverified and check a current developer update before relying on it. It should be kept attached to the main product rather than blended with an earlier preview or demo limitation. The research does not provide a verified name and full ability description for each of those 18 entries.
A complete-looking list assembled from scattered older posts could therefore mix development stages. This page keeps the confirmed total while avoiding an invented current roster. This keeps unrelated editor choices from obscuring what changed. It also gives you a clearer basis for describing a reproducible result to other players, without turning one observation into a guaranteed rule for every save or build.
Why do older pages say five classes?
The original demo was released in June 2026 with a smaller content set. The official material establishes the documented feature, while the dated gameplay examples help show how that feature appeared in a particular build. Keeping those two evidence types together makes it easier to separate a confirmed mechanic from an assumption based on another game. Its description included five classes alongside limited environmental and settlement themes.
A difference between those versions can explain apparently conflicting instructions without proving that one source is universally correct. If a step or value is not demonstrated in the collected sources, treat it as unverified and check a current developer update before relying on it. That was a demo scope statement, not evidence that the later main game had only five classes. The demo was later withdrawn pending a rework.
A future reworked demo would need its own verified content description rather than automatically inheriting either the old five-class limit or the main game's 18-class count. Historical availability and content limits should remain dated. This keeps unrelated editor choices from obscuring what changed. It also gives you a clearer basis for describing a reproducible result to other players, without turning one observation into a guaranteed rule for every save or build.
Are professions the same as classes?
The collected materials distinguish hero or combat-related class concepts from civilian professions such as farmer and miner. The official material establishes the documented feature, while the dated gameplay examples help show how that feature appeared in a particular build. Keeping those two evidence types together makes it easier to separate a confirmed mechanic from an assumption based on another game. A resident's economic role should not automatically be added to a combat-class roster.
A difference between those versions can explain apparently conflicting instructions without proving that one source is universally correct. If a step or value is not demonstrated in the collected sources, treat it as unverified and check a current developer update before relying on it. Combining the two would inflate the list and make class comparisons misleading. The broader population setup also uses civilian and soldier templates linked through factions to City Zones.
Those templates establish which inhabitants can appear in a settlement. They are related to how a world is configured, but the collected sources do not establish that every template type is a separate class. This keeps unrelated editor choices from obscuring what changed. It also gives you a clearer basis for describing a reproducible result to other players, without turning one observation into a guaranteed rule for every save or build.
Which individual examples are safe to discuss?
An October preview includes a Paladin example. The official material establishes the documented feature, while the dated gameplay examples help show how that feature appeared in a particular build. Keeping those two evidence types together makes it easier to separate a confirmed mechanic from an assumption based on another game. That supports describing it as an observed example in that recording rather than treating it as a complete roster or a recommended best class.
A difference between those versions can explain apparently conflicting instructions without proving that one source is universally correct. If a step or value is not demonstrated in the collected sources, treat it as unverified and check a current developer update before relying on it. The footage does not establish a full comparison of every class under equal conditions. Older development posts also contain names from earlier stages of the game's class design.
Without a verified current roster, those historical names should not be promoted into an exhaustive launch list. An example's existence in old material is not enough to establish its present abilities, balance or availability. This keeps unrelated editor choices from obscuring what changed. It also gives you a clearer basis for describing a reproducible result to other players, without turning one observation into a guaranteed rule for every save or build.
Can the evidence support a tier list?
A useful tier list would require verified current classes and meaningful comparative evidence. The official material establishes the documented feature, while the dated gameplay examples help show how that feature appeared in a particular build. Keeping those two evidence types together makes it easier to separate a confirmed mechanic from an assumption based on another game. The collected research does not provide that combination.
A difference between those versions can explain apparently conflicting instructions without proving that one source is universally correct. If a step or value is not demonstrated in the collected sources, treat it as unverified and check a current developer update before relying on it. It therefore cannot establish the best starter class, the highest damage option or a universal ranking for different playstyles. The same limit applies to detailed unlock recipes and skill statistics.
No supported universal class-selection sequence or numerical performance table is established here. Keeping those claims out is more useful than presenting an authoritative-looking ranking built from mismatched demo and preview information. This keeps unrelated editor choices from obscuring what changed. It also gives you a clearer basis for describing a reproducible result to other players, without turning one observation into a guaranteed rule for every save or build.
How do classes relate to character control?
Creature control is advertised, but the collected footage does not verify a successful current-version possession workflow. The official material establishes the documented feature, while the dated gameplay examples help show how that feature appeared in a particular build. Keeping those two evidence types together makes it easier to separate a confirmed mechanic from an assumption based on another game. A historical demo attempt failed with a missing linked simulation hero, which does not establish that choosing a particular class always fixes the problem.
A difference between those versions can explain apparently conflicting instructions without proving that one source is universally correct. If a step or value is not demonstrated in the collected sources, treat it as unverified and check a current developer update before relying on it. Class configuration and direct control should remain separate questions unless a source demonstrates their exact relationship. For worldbuilding, learn the faction, NPC-template and City Zone dependencies before trying to infer a class's role from appearance alone.
For current class details, compare the official main-game description with build-specific developer updates. The accurate takeaway from this material is the advertised 18-class main-game total, the historical five-class demo scope and the absence of a verified complete current ranking. This keeps unrelated editor choices from obscuring what changed. It also gives you a clearer basis for describing a reproducible result to other players, without turning one observation into a guaranteed rule for every save or build.
Sources and verification
This answer cross-checks developer and Steam material against independent reporting or gameplay footage. Details from older demos are identified as historical because interfaces and availability can change between builds.