Your race decides a real starting constraint before the full Skills tab is available. Do not pick from the largest green number alone. Choose the penalty your first build can absorb, then use natural skill growth before spending the five manual points awarded on a level-up. That order matters because a point spent now raises the level that later action-based gains must climb from.
This guide reflects the Early Access client and community guide state checked on August 19, 2026, after official Hotfix 4. NerveLabs is updating the game rapidly, and the character guide itself warns that its values may change. Treat every number below as a current starting-screen value: if your client disagrees, the live card wins.
Begin with the penalty that can break your intended loop
The five starts do not use the same kind of trade. Umbrite has three positive skill adjustments and no listed numerical drawback. Al-Fijhani buys trade and lock access with health. Ruggite accepts both a stamina loss and a fire weakness for a much larger magic pool. Thrakh trades archery for carrying and swimming, while Calyx trades armor and swimming for Psionics and Night Eye.
That means the first question is not “Which bonus is strongest?” It is “Which listed cost will interfere with the actions I expect to repeat?” A player who wants bows should notice Thrakh’s Archery deficit before the carry bonus. A close-range player should price Al-Fijhani’s health loss before celebrating cheaper purchases. A caster still needs a plan for movement and fire damage if choosing Ruggite.
| Starting race | Visible adjustments in the checked build | What the package naturally supports | What must be managed |
|---|---|---|---|
| Umbrite | +5 One-Handed; +5 Two-Handed; +5 Channeling | An opening centered on a chosen melee weapon, with Channeling kept as an available third lane | No listed numerical penalty, but splitting use between both weapon skills can dilute early practice |
| Al-Fijhani | +15% Better Prices; +15 Lockpicking; -20 Max Health | Trading, lock access, and a route that profits from returning to settlements | The health loss makes careless close-range exchanges more expensive |
| Ruggite | +100 Max Magic Power; -30 Max Stamina; 50% Weakness to Fire; +1 Starting Spell | A magic-heavy opening with a much larger starting resource pool | Lower stamina and a large elemental weakness require an escape and positioning plan |
| Thrakh | +50 Max Carry Weight; +15 Swimming; -10 Archery | Long collection trips, heavy inventories, and water routes | The Archery deficit conflicts directly with a bow-first plan |
| Calyx | +15 Psionics; -5 Armor Skill; -10 Swimming; +1 Spell (Night Eye) | Psionics and early visibility support without needing to find the first utility spell | Armor and swimming begin behind the neutral baseline |
The “supports” column is a planning inference, not a hidden class restriction. The game lets skills rise through use and accepts manual allocation, so none of these starts proves that another playstyle is impossible. It only tells you where the first hours will have less friction.
Use each race as a starting package, not a finished build
Umbrite: commit to one weapon before training the other
Umbrite begins with bonuses to both One-Handed and Two-Handed. Carrying both does not mean both need equal use. Pick the moveset and reach that solve your immediate fights, watch that skill’s progress, and leave the other weapon type as a later branch. Channeling is also boosted, but the checked evidence does not establish a mandatory interaction between Channeling and either melee skill. Do not invent one just to use every bonus immediately.
Al-Fijhani: turn access and prices into a survival plan
Al-Fijhani’s Better Prices and Lockpicking bonuses both reward interacting with the world rather than forcing every encounter. The -20 Max Health is the counterweight. Before spending points on damage, check whether the current problem is actually failed access, poor inventory value, or taking one hit too many. If health is the failure signal, a trade bonus does not compensate inside the fight; change position, equipment, or approach first.
Ruggite: budget around two liabilities, not one large pool
The +100 Max Magic Power is the most visually dramatic starting increase in the current list, but it arrives with -30 Max Stamina and 50% Weakness to Fire. A larger pool lets you attempt more magic actions; it does not guarantee that a spell is efficient, that stamina will recover at the right moment, or that fire can be ignored. Keep a non-fire threat as your first test target, learn the starting spell before crafting a wider rotation, and preserve enough stamina to disengage.
Thrakh: exploit capacity without accidentally choosing an archer
Thrakh’s +50 Max Carry Weight makes the benefit easy to verify: compare the inventory limit before deciding what to keep. The +15 Swimming supports water travel, while -10 Archery is a real starting deficit. If the plan is bow-first, that conflict should be deliberate rather than discovered after the opening. If the plan is hauling, melee, or exploration, the package has a clearer internal logic.
Calyx: separate visibility from durability
Calyx begins with +15 Psionics and Night Eye, but also -5 Armor Skill and -10 Swimming. Night Eye solves a visibility problem; it does not repair the armor deficit or make water travel neutral. Put the spell on a reachable slot, verify its current cost and behavior in a safe dark area, and judge defense from actual damage taken rather than from the convenience of seeing the attacker sooner.
Let repeatable actions work before spending the five points
The current Skills tab supports two advancement paths. Related actions fill a progress bar toward the next single skill level, while a level-up grants five points that can be assigned with the plus control. The important consequence is timing: Sprinting, Swimming, and Sneaking can rise while you move through normal play, but a difficult combat or utility skill may offer fewer safe repetitions.
- Open the Skills tab immediately after character creation. Record the actual levels and any negative values. The Calyx example in the checked guide displayed Swimming at -10 rather than silently clamping it to zero.
- Choose one action you will repeat anyway. Sprint on a normal route, swim only where the return is safe, or sneak through an area where detection will not trap the run. Watch for progress rather than assuming every animation awards experience.
- Leave the five points unspent long enough to identify the slow lane. If movement advances during travel but the chosen combat skill barely moves, the manual points have a clearer job.
- Spend against a blocker, not a fantasy end build. Add enough to improve the action that is preventing progress, then return to the Skills tab and read the new level and progress requirement.
- Recheck after the next natural gain. The current community evidence says higher levels require more natural experience. Spending all five first can therefore make the next action-earned increase slower.
This is why “always spend points immediately” is a poor Early Access rule. The five points are flexible; natural practice is also progression. Use the easy repetitions first, then apply points where repetition is risky, slow, or unavailable.
The Skills tab is the build audit
A guide can tell you what was visible on August 19, but only the current client can prove what your save received. Use four readbacks:
- Current level: confirms that the racial adjustment and any manual point were applied.
- Progress bar: shows whether the action you are repeating is actually training the intended skill.
- Selected-skill description: prevents a familiar-sounding skill name from being assigned a guessed effect.
- Plus control and available points: confirms how many points remain before another allocation.
If a sprinting session moves the wrong bar, stop and inspect the action state. If a racial number differs from this table, capture the live card and use the client value. If a spell or weapon action produces no visible progress, do not spend all five points to hide the uncertainty; first confirm that the skill governs that action.
Three opening plans that respect both sides of the choice
| Player goal | Coherent start | First verification | Reason to switch |
|---|---|---|---|
| Simple melee learning run | Umbrite with either One-Handed or Two-Handed as the active lane | One weapon’s bar rises during real fights | You actually want trade, magic, carrying, or Psionics to define the run |
| Exploration and economy | Al-Fijhani for locks and prices, or Thrakh for carry weight and swimming | The intended access or capacity benefit changes a real route | Al-Fijhani health or Thrakh Archery penalty undermines the planned combat style |
| Magic-first experiment | Ruggite for pool size, or Calyx for Psionics and Night Eye | The starting spell and relevant skill can be used and tracked safely | Ruggite’s stamina/fire costs or Calyx’s armor/swimming deficits dominate play |
These are decision frameworks, not tier rankings. They deliberately avoid permanent claims about damage, spell costs, recovery, or a universal best race because Hotfix 4 arrived less than two days after launch and official updates are still rapid.
Freeze the decision with a one-minute client check
Before investing a longer run, read the character card, create the character, open Skills, and confirm every number that matters to the plan. Write down the five manual points after the first level-up, let one easy skill gain naturally, and only then allocate. That sequence distinguishes a real build advantage from a bonus that looked attractive but never served the actions you use.
Once the opening build is stable, the Nuisance quest housing route is an early practical test: it includes road travel, two skeleton encounters, a quest item, and a navigation clue without requiring a broad Tower build.
Sources & credits: release state, AppID 3849000, Early Access status, Hotfix boundary, and the published image were checked against the official Steam record. Starting adjustments and Skills-tab behavior were verified against the current English character-creation community guide and its visible August 19 update. Disputed race flavor text was excluded.