ClassesUnearthed Arcana

Unearthed Arcana: Artificer Breakdown

Mearls & co. kick the new year off with a thunderous roar in today’s new UA article. Now, one of these days I’ll cover the artificer in the History of the Classes, but the short version is that the first serious artificer class shows up in Eberron, as part of 3.5e, and again in 4e. Much as the mystic class seemed to signal that they hadn’t forgotten about Dark Sun, this is a reminder that they haven’t forgotten Eberron (which, of course, also needs mystics to function), and they’re open to making 5e more steampunk-friendly. (Also, quick plug for my esteemed colleague’s Alchemist class – but if you’re a frequent Tribality reader, you already know how great it is.)


I think I’ve talked before about how some 5e classes put most of their mechanical weight in the core, so that the subclass is a peripheral addition, while other classes are the reverse – a nearly empty container that unites mechanically weighty subclasses. For example, any two barbarians have a lot more in common than any two warlocks, right? Well, the artificer’s core features offer little on their combat style and gameplay feel.

  • d8 hit dice
  • Light and medium armor, simple weapons, no shields
  • Thieves’ tools and two other tools of your choice
    • Since the class otherwise emphasizes Dex and Int, that’s a good basis for being the party’s traps rogue.
  • Con and Int saving throw proficiency
  • Three skills from a list that suggests a roguish scholar
  • Detect magic and identify as rituals, without material components. Definitely on-message here.
  • Tool Expertise – double your proficiency bonus in the three tool proficiencies you gain from this class.
  • At 2nd, 5th, 10th, 15th, and 20th level, you gain a wondrous magic item, based on off-camera research and crafting. The list of choices here are pretty good, and heavily slanted toward utility and movement rather than pure combat value. This list can’t help but look thin as the edition matures and more magic items enter the metagame, though.
    • Guaranteed magic items exist in a really odd tension with the foundational decision to never, ever assume the PCs have magic items. Also, I think this is the first time we’ve seen class features that could be stolen and can’t be replaced with a standard investment of time, except insofar as you could try to use the miserable magic item crafting rules of the DMG. I’ll come back to the issues of 5e’s crafting and research mechanics, though.
  • Spellcasting, progressing like Eldritch Knights or Arcane Tricksters – 1st-level spells starting at 3rd level, and progressing to 4th-level spells from there. The spell list is full of defense, mobility, and utility; a few damage-buffing spells (like magic weapon and enlarge), but no direct damage at all. Looks like a particularly heavy reliance on Concentration, so it’s good that Con is a proficient saving throw. Intelligence is their casting stat, but they have a set number of spells known, rather than a spellbook or knowing all spells of the levels they have access to.
  • Infuse Magic lets you store your spells in items (for up to 8 hours) and hand them off to your allies. You can store a number of spells at a time equal to your Intelligence modifier.
    • It’s pretty cool; the one thing this doesn’t clarify is whose Concentration the stored spell requires. Getting to shift the Concentration burden onto your non-casting allies is incredibly good; not getting to shift Concentration to your non-casting allies means they might accidentally interrupt one of your spells, or each other’s spells.
  • Five Ability Score Improvement/feat slots.
  • You get to attune more magic items – one extra slot at 5th level (are you really likely to have four items that need attunement by this level? Well, thanks to Wondrous Invention, yes), and a second extra slot at 15th. (There’s a sixth slot coming, but it’s a separate feature, so I’ll get to it in a second.)
    • This is fine, but again strikes at the foundation of “we don’t assume you have magic items.” This is a lot of lost power if you don’t have magic items!
  • Mechanical Servant gives you a construct pet in beast form, a modification of a beast up to CR 2. (Lifting a joke from a 5e Facebook community: this obviously makes its name CR2-D2.) I’m not sure if it’s the best (I haven’t done a full comparison), but a saber-toothed tiger golem sounds pretty badass to me! Rhinoceros and polar bear are also top-flight options. Anyway, it gets a nice set of template upgrades from being a construct, but it doesn’t scale with level at all.
    • This feels like a weird core class feature, especially because of the hullabaloo around the Beast Master ranger. It doesn’t cost anything from the artificer’s action economy, and it is a huge spike in the artificer’s power for a short time, but its relevance falls off over time, so it’s less clear whether this is supposed to be a central piece of the artificer’s combat style. I strongly suspect this will get rewritten before final publication.
  • Soul of Artifice gives you your sixth attunement slot, and you add your current number of attuned items to all saving throws you roll. It’s going to be real hard to fail a Constitution or Intelligence saving throw now! (d20 + Con/Int mod + 6 for proficiency + 6 for Soul of Artifice).

Up to this point, this is a strange little class that doesn’t do anything all that well except play defense, and relies on its pet for damage output. That is about to change, because the two subclasses are the Alchemist and the Gunsmith, and they are 100% about blowing things to the Nine Hells.



The fundamental problem with an alchemist class is that rules for crafting in tabletop games have never been very good, but mixing up potions from components is the defining idea of an alchemist. Also, the ability to turn cash into extra power has serious problems on both the Too Much Cash and Too Little Cash sides, and D&D really doesn’t want to tie the DM’s hands by making enough cash handout the core of whether or not a class is fun to play. (I’ve fought with this design problem in tabletop and live-action games for the better part of the last ten years. It ain’t easy.)

The solution WotC goes with is to handwave crafting by making it happen in a magical satchel. I think it misses a lot of the compelling fiction of playing an alchemist, much as I think a wizard who doesn’t get to research spells and accumulate a library is missing a lot of the fun part of playing a wizard.

  • The Alchemist’s Satchel is a necessary focus item that can, in theory, be stolen or destroyed, but at least it can be reconstructed. All of your Alchemical Formula options automagically come out of this.
    • It reminds me of the dwarves in Myth II: Soulblighter, so that’s a plus.
  • You start with three Alchemical Formulas: Alchemical Fire, Alchemical Acid, and one other from a list of five. You learn four more (leaving one forever unknown) over the course of your progression. Learning new formulas and scaling up the effects of these formulas is all you get for the rest of this subclass’s progression.
    • Alchemical Fire is an at-will 5-foot-radius firebomb, which starts at 1d6 and eventually scales to 7d6. Its 30-foot range is a pretty serious limit, at least.
    • Alchemical Acid is an at-will single-target acid attack, which starts at 1d6 and eventually scales to 10d6. It also has a 30-foot range limit.
    • Healing Draught is a once-per-target-per-their-long-rest (it’s sort of odd timing, but sensible for balance) heal, initially small and unreliable with 1d8, but scaling up to 10d8. In combination with cure wounds, lesser restoration, and revivify on the artificer class list, an alchemist should be a fairly competitive healer or backup healer. This is even written to mitigate action-economy issues, as you can hand out one healing draught before combat, to put the action cost on the person healed.
    • Smoke Stick is a fog cloud, more or less – a heavy obscurement that blocks Devil’s Sight and similar superior-darkvision effects (because it’s smoke). Slightly smaller area than fog cloud. Anyway, I love seeing this kind of thing used well, but it doesn’t come up all that often. It’s limited to once per minute.
    • Swift Step Draught is a double-strength longstrider with a 1-minute duration. It’s not always what you need, but at just a bonus action to pull it from the satchel, it’s a great trick. Also limited to once per minute.
    • Tanglefoot Bag makes a 5-ft-radius difficult terrain for 1 minute, and it’s a 50% speed snare if they start their turn in that area. Also limited to once per minute.
    • Thunderstone is an at-will 10-foot-radius knockdown and 10-foot knockback. No damage, Constitution saving throw to resist.

That’s the Alchemist. It bothers me that all mixing and research are handwaved, and there’s nothing to learn outside of what the level progression automatically grants. Their damage output is pretty scary, though 10d6 per round is still less than an eldritch blast with Agonizing Blast and a 20 Charisma (10d6 averages 35, 4d10 + 20 averages 41). Note that the Dex saves against Alchemical Acid and Fire are not for half damage – on a success the target takes no damage, full stop. Targets with Dex save proficiency will be a bad problem, but this is not super common with monsters.

Ultimately, the Alchemist playstyle has more in common with the 3.5 warlock than anything else – Alchemical Acid and Alchemical Fire wind up resembling the 3.5 eldritch blast more than a little, and they’re backed up with fairly similar additional tricks. Oddly enough, the Gunsmith subclass is even more like the 3.5 warlock, as I am about to explain.

To make myself less unhappy with the absence of mixing and research rules, I’ll eventually create an alchemical side effects table that the Alchemist can roll on as part of any long rest or downtime. It would have a mix of good, bad, and weird ongoing effects. ETA: “Eventually” turned out to me “tomorrow.” Here’s my Harbinger post of alchemical experimentation.



As you might expect, the inclusion of guns – even magic guns – is hugely divisive in online discussion about this document. Mearls has pointed out that the gunsmith got shoved into a subclass rather than being a primary class, so that you lose less content overall if the DM decides guns don’t fit the campaign. Anyway, this subclass is 100% about the Thunder Cannon (uber firearm), making it better, and adding more options.

  • Master Smith grants another tool proficiency and the mending
  • I think this is one of the first times a subclass feature has actually given you a unique weapon. It could nominally be enchanted, but it’s not completely clear whether you can expect to see magical Thunder Cannons as treasure, nor what it takes to use a Thunder Cannon that you did not make. (Let’s say you’re not a Gunsmith, but there’s a Gunsmith in your party…) Anyway, it’s a firearm that does not have the Loading property per se, so you can’t get around the Loading property with the Crossbow Expert. (The same effect, but without the keyword. I’m sure this will never confuse anyone!) Anyway, there are rules for making a new Thunder Cannon if you lose yours.
  • Arcane Magazine makes sure you have to track ammunition, but don’t have to work very hard at it. I’m sure no one will drive their DMs completely berserk with the obvious effects of having an unlimited supply of gunpowder. (But having to purchase gunpowder doesn’t help all that much – if it’s available enough for gameplay to proceed, it’s available enough for That Dude to expect the DM to let him make bombs.)
    • Anyway, the Arcane Magazine creates 40 rounds at the end of a long rest, and 10 rounds at the end of a short rest. I’m not really sure why these rounds aren’t self-destroying-if-unused the way alchemical creations are, but that’s probably a necessary addition.
  • Features up to this point are all 1st-level features.
  • Thunder Monger makes your Thunder Cannon’s damage scale up, +1d6 thunder damage per two levels.
    • Thunder Monger is an action in itself, not part of the Attack action, so things that trigger with the Attack action are disconnected from this. I’m not sure what those things would be, but it’s future-proofing. The damage scaling of Thunder Monger is also detached from the scaling of the next couple of features. However, it also opens the door to the designers getting tripped up in 5e’s technical verbiage around attacks.
  • Blast Wave is the classic blunderbuss blast – area force damage and a 10-foot push. The verbiage here does get weird, as it is a “special attack” that doesn’t make an attack roll. (With only a tiny number of exceptions, attacks involve attack rolls rather than forcing saving throws.) It does okay damage, but this is mostly about battlefield control. Also, the saving throw is Int-based, so Gunsmiths actually need Dexterity and Intelligence, where an Alchemist can get by with just Intelligence.
  • Piercing Round is a 30-foot line attack, because every game eventually lets you shoot straight through your targets. Much like Blast Wave, this is a special attack that forces a saving throw. I’m not completely clear on whether special attacks require ammo. The damage scales from 4d6 to 6d6 lightning damage.
  • Explosive Round does what it says on the tin: a special attack that deals 4d8 fire damage in a 30-foot radius.
    • Note that Blast Wave, Piercing Round, and Explosive Round are all no-effect-on-save, but at-will. So it’s incredibly important to push your Int-based saving throw DCs as high as they go; in a sense you’ll need to decide whether you’re speccing for Thunder Monger or AoEs, because you might eventually have a 20 in both, but it’ll probably eat up some or all of your feat slots to get there.

And that’s all for the Gunsmith. The four different attack options – single-target, cone, line, and radius – are what make this feel like the 3.5 warlock with its different at-will AoE options to me. I like that they get five different flavors of damage, but the existing firearm proficiency rules are… awkward here, because it just takes the PCs spending time so that every team member can use the Thunder Cannon.

The Alchemist and the Gunsmith both presume a high integration of magic and technology, which is…  not for all settings. The classic Arcanum: Of Steamworks and Magick Obscura is not on board with this approach, for example. It’s Eberron-friendly, in my view – they have lightning trains and magic fiery airships, so thundering longarms don’t feel like a big stretch to me. Mearls and Crawford are going for the just-get-to-the-action approach that we see in a lot of video games, minimizing resource management. The 4e artificer’s healing concept had more believability going for it, as it was a transmutation worked on someone’s blood (well, that was our interpretation!), rather than having nigh-infinite healing potions in your back pocket. An Alchemist could heal an army, at one soldier per round.

In final consideration, I am glad that the artificer class exists. I don’t think its design is all the way there yet, but that’s why we playtest. I would allow and encourage the homebrewed alchemist in my own campaign to give this some playtesting time, if she wanted. (Remember how I thought my second child was about to be born? Still waiting. Today is his actual, 40-week due date.)