I've got a lot to go over so I'm going to break it down by level. I do a lot of homebrewing (including full classes), so I will walk you through some of the balance checks I do on my own work (usually I build a class based purely on flavor and what sounds cool and then bring it into line with the Player's Handbook classes after).

Comparison Metric

Based on the progression we see, we will be comparing to other full casters for the most part. With 1,6,14,20 subclasses we are looking to the Cleric subclass progression with some delayed high-end progression (cleric is 1,6,8,18).

The main class features read a lot like the Bard with small Rogue-like ports so I'll be comparing them with the Bard features.

Level 0 (Proficiencies and the like)

The proficiencies for this class are fairly normal. They are very similar to the Bard's concept in sacrificing weapon/armor options for an additional skill and tool proficiency.

The saves are also largely fine (Dexterity is the most common save while Intelligence is the least).

Nível 1

Feitiço

This is essentially balanced modeled on the Sorcerer progression. More cantrips than the other classes but limited spell selection options and less spells known. This is fine as long as the spell lists is na realidade limiting. The way it is worded is ambiguous as many spells could be seen as "enhancing your own abilities... or some kind of utility for a skulk" (ênfase mine). I recommend telling your player to make an actual list and make sure it as limited in scope as the sorcerer spell list.

Beguiling Style Level 1

The Mindbender is fairly comparable to the Knowledge Domain cleric at level 1. It trades the second expertise and two languages for telepathy which is fine. The skills for the Mindbender are more broadly useful than the ones for the Knowledge domain, but the useful of any skill is contingent on the GM so mileage may vary. In general, the "conversation" skills are not particularly more potent than the "knowledge" skills.

The Shadowcrafter, however, is a problem. The level 1 part of Shadow Crafting is fine. Gaining Minor Illusion and a Minor Creation type effect is fairly comparable to the 1st level of the Arcana Domain. Creating a 1d8 weapon is comparable to the Shillelagh cantrip (although it can make ranged weapons too, it doesn't change the modifier), so you are virtually getting two cantrips like the Arcana domain. The Darkvision could be seen as a parallel to the Arcana proficiency (may sometimes overlap anyway, and is a fairly minor benefit). So if the Shadowcrafter apenas got Shadow Crafting and Darkvision that doesn't scale than it would be balanced.

Instead, the Shadow Crafting faz scale and becomes extremely unbalanced. Cantrips are able to scale like this because they require the Cast a Spell action. Instead, a weapon uses the Attack action which can be facilitated by many things such as the pressa spell or the Extra Attack feature. Shadow Crafting becomes almost as good as the 2nd level spell lâmina de sombra at level 5. Also, the creation process necessidades to be outline. Is it a bonus action like Shillelagh? Does it require a period of rest before it can be done again? Does it take a period of downtime? Shadow Crafting needs to either not scale or be restricted to once per action like a cantrip.

Side note: the other equipment Shadow Crafting creates needs to be specified that it could not be used as Spellcasting components. Creating other types of equipment in general is problematic at it is so versatile. It is shared by the Forge Domain's channel divinity, but the cleric has to actually pay the costs. I would restrict it to the weapon only

E, então there is Shadow Cloak. If the class apenas gave Shadow Cloak at level 1 it would be overpowered. First of all, how is this activated? An action? Bonus action? What is the Perception DC? Do they get Disadvantage even if you don't use Stealth to hide? Next, the daylight restriction is not enough to balance it. Restrictions like this just mean the party can play around the day and requires you to force them to fight during the day to negate this feature or just let it run wild. Finally, the details. This feature is as good if not better than the Invoke Duplicity Channel Divinity of the cleric. Invoke Duplicity grants advantage which is usually better than disadvantage to enemies, but it is limited to once per day (a couple times at higher levels). Shadow Cloak needs to be either removed or limited severely. Move it to level 2 and limit it to once a day (more at higher levels) and it becomes like Channel Divinity. Then give Mindbender something else limited at level 2 to balance it with Shadowcrafter

Nível 2

Perícia

Since the main class features are being compared to Bard, this is largely equivalent. The first round of expertise comes in one level earlier, but the second round comes in one level later. This is fine.

Hidden Casting Level 3

Since the main class features are being compared to Bard, I'll include the usefulness of this next to Jack of All Trades and Song of Rest (as Expertise aligned with Channel Divinity). The subtle spell effect is weird.

These effects are only applied when the target has an enemy other than you within 5 feet and that enemy is not incapacitated, or when the target has disadvantage.

What does this mean when you cast a spell that doesn't have a target. When the target has disadvantage on what? These need to be elaborated on. If it means when the target has disadvantage on saving throws against your spells, then this will almost never come up. If it means, "when the target would have disadvantage on an attack against you," then it is probably on par with the Bard's two similar features: a minor benefit that appears in many situations. However, why take that niche from sorcerers? Being able to ignore components is part of the sorcerer's thing. I would ask that the player change this feature to something equally minor but different.

4th level

Ability Score Increases should also occur at level 8, 12, 16, and 19. I'm going to assume this was a typo and leave it there saying this is standard.

Beguiling Style Level 6

Controlador de mentes

You learn a free spell in sugestão but with a longer duration. Applying a modifier like this to a spell is not unheard of (it appears in some Eldritch Invocations), but is a little strange. Another ambiguity: Is it 20 minutes per level of spell slot used, or 20 minutes per beguiler level? This is probably on par with the benefits of a cleric's domain. Consider Life Domain with the feature Blessed Healer. A small improvement to a handful of spells.

E então there is the levelled portion of Mentalist. This just needs to go. Reading surface thoughts is something reserved for a second level spell for a reason. It shouldn't just be always on. Make the character have to actually cast the spell. Perhaps instead it should just add the spell detectar pensamentos to the list of spells you can learn (kind of like Domain spells).

Shadowcrafter

Shadow Illusion involves more math than 5e is meant to have (5e tried to make the game smoother). As for the substance of it, replicating spells is a problem. The Spellcasting progression for the beguiler was based around the Sorcerer for it's limitado spell list. Giving access to these spells, even at the cost of a higher spell slot, is not a good idea. Trading spell slot level for versatility (in the same vein as a sorcerer's Metamagic trades sorcery points for versatility), is very strong. Instead, add a few spells to the spell list like domain spells but make them illusion instead for the beguiler.

I'm not sure what the Shadow Crafting improvement means as as discussed earlier, the Shadow Crafting feature needs fixing. Conjuring small monsters, however, should be the realm of conjuration spells (such as I discussed with the spell portion of Shadow Illusion).

The illusion-identifying element is weird. It seems to only apply to the spell portion of Shadow Illusion (the first paragraph) but it appears at the end. Also, what is "shadow material." If it is widely available this restriction becomes largely meaningless. I would just remove this entirely in favor of an increased spell access (see the next section).

Domain Spells

If you give both subclasses extra spells, model it off of the domain spells. The mindbender would gain spells at each of 5 levels with small changes (perhaps change damage type to psychic for some) and the Shadowcrafter would gain spells at each of 5 levels with small changes (the illusion school and perhaps some damage type changes with illusion-identifying elements).

Hidden Casting Level 9

This feature is ridiculously strong. It is comparable to Shadow Sorcerer's Hound of Ill Omen, but you can't end the effect by killing the Hound. Sure it comes in later than the Hound does (and the Hound causes disadvantage which is more potent), but as with the level 3 portion of Hidden Casting, the enemy nearby is too easy to achieve. Reduce this to a +1 and it is still a bit too strong, but pairing it with the level 10 comparison (see below) makes it okay.

Nível 10

Evasion is fine. It is being compared to Magical Secrets. Bard's Magical Secrets is one of the best features in the game as it allows for some crazy synergies so Evasion and the a weakened Hidden Casting level 9 feature makes it fair.

Beguiling Style Level 14

Controlador de mentes

Permanent charming just doesn't fit. If this is done on one of your main threats in the game, it is story-breaking as that enemy is now permanently subdued. If it is done on an underling, it is functionally useless as all the Charmed condition says is...

A charmed creature can’t attack the charmer or target the charmer with harmful abilities or magical effects.

... and grants advantage on social checks (in which temporary charming is usually good enough). It is a too weak when compared to the cleric benefits of either Divine Strike or Potent Spellcasting which come out 6 levels earlier. Find something that is more widely applicable and give it some kick

Shadowcrafter

Control what illusions? The monsters from Shadow Illusion? You never said it required an action to control. Do you mean any monster from a Conjuration spell that is replicated by Shadow Illusion, because that is not an aspect of Conjuration spells. I don't know what this feature means, so I will ignore it for now. **You want to incorporate a feature with a slightly higher power level than and which is as frequently useful as Divine Strike or Potent Casting.

Fixing level 14

Since the cleric gets a widely applicable feature at this spot in the subclass progression, I would do the same for the beguiler. Give the subclasses a benefit on their cantrips and "domain spells" I discussed earlier. I would have the Mindbender cause a circumstantial disadvantage to saving throws for these spells and have the Shadowcrafter gain an increase in damage equal to Intelligence. For the Mindbender's disadvantage, I would use "when you deal psychic damage to a creature with a spell you cast, it has disadvantage on saving throws against your spells until the end of your next turn."

Hidden Casting Level 15

We are comparing this to the next round of Bard's Magical Secrets. At this level, Magical Secrets is not nearly as useful as before as higher level spells that a bard would be chosen are more limited by the number of spell slots available, but it is still a fairly potent feature to compare to. So what does the beguiler get? ...

I have honestly know idea what this means. Does this essentially add 12 seconds to the duration of all spells that require Concentration, even after concentration is broken!? If so, this ability is maneira stronger than the Magical Secrets improvement. The primary balancing act of Concentration spells is that they can be broken by dealing damage to the caster. Combat in 5e rarely lasts more than a few rounds so this basically means a concentration spell will always finish out the battle. If the beguiler needs to have some sort of benefit to Concentration spells, give them a bonus (perhaps advantage) to concentration saves instead (one such official example of a feature like this is the War Magic wizard subclass).

Nível 17

This feature is being compared to the level 18 Magical Secrets improvement. The bard's extremely potent feature returns to its former glory from its weaker state at level 14 in this form as it grants access to all of the powerful (sometimes game breaking) 8th and 9th level spells in the game. And we compare to... Slippery Mind? Okay so clearly, this is muito weaker. Let's see if its made up for at level 19 (read on).

Nível 18

Assim...

At 1st level, pick a Beguiling Style to focus on how you will misdirect and mystify your opponents.

Your choice grants you features when you choose it at 1st level and again at 6th, 14th, and 18th level.

[emphasis mine]

But there are no 18th level features in the Beguiling Styles so I'm going to assume for now that the level 20 features are supposed to appear at level 18.

Controlador de mentes

Thrall is essentially a remolded version of the Necromancy Wizard's Command Undead. At level 18, it can be a little bit more powerful, so let's see what we have.

Command Undead is limited by Intelligence while Thrall is limited by Challenge Rating. This makes Command Undead much more situational as the persistent save that Intelligent undead get (most high CR undead have higher Intelligence) makes the feature muito weak in these situations. Thrall is more broadly applicable. Thrall is ainda too strong since it works on any creature of CR 13-15 para sempre. Você pode either make it half of beguiler level (instead of minus 5) to allow for a permanent weak companion (what Command Undead often results in) or provide a limit on what powerful monsters can be chosen (Command Undead does Intelligence as a limiter for example). This feature should be a bit stronger than Command Undead, but not by much. In the cleric field it is comparable to equally powerful features like the Grave Domain's Keeper of Souls.

Shadowcrafter

Like with Shadow Illusion, the math here doesn't fit in with 5e and regardless, the effective is maneira too variable. The spells created by Spell Illusion could deal a little bit more damage or double damage. Like with Easy Concentration, I'm unsure what creatures this feature is referring to. If you took my advice for using an ability like Potent Cantrip for level 14, then the spell illusions are already more powerful so you could focus entirely on the creatures. Simply making the creatures "more real" would be similar to the Conjuration Wizard's Durable Summons and like with the Thrall feature, it can be a little bit stronger at the higher level. One solution would be to grant resistance to all damage to your summons (maybe with "except Radiant") as the shadow material is strengthened from deeper recesses of the Shadowfell.

Hidden Casting Level 20

The "fun" capstone is really coming from the Beguiling Style (similar to Cleric) so the Hidden Casting should be pretty secondary. Instead we get...

you overcome all spell resistance

... I have only a tiny idea what this means. I think it is going for something like:

Creatures gain no benefit from damage resistances or features that grant advantage on saving throws against spells you cast.

If that is the case, then this works fine as a secondary capstone.

Conclusão

The biggest problems come from poor wording. Design features so they work well with 5e. Shadowcrafter and Hidden Casting are either a little weak or very muito strong depending on how you interpret some of the features. Mindbender is the most balanced, but also has some wording issues. The "core" of the class is largely fine as long as you fix the subclasses and Hidden Casting.

Level 6 and level 14 are the biggest problem areas. I would entirely rewrite level 6 as the features are just too weird. 5e does not have spell-like abilities, instead favoring abilities that just cast the spells for other resources (see the Way of the Four Elements Monk). Shadowcrafter Level 14 just doesn't make sense and Mindbender is very weak and doesn't fit in with how the game is usually played. I would change these entirely.

Finally, this class is very interesting and has some downright awesome but suffers from three of the most common pitfalls for homebrew. One, some features are poorly thought out and improperly worded. Two, the "coolest" (i.e. most complex) features aren't checked for balance. And three, the features incorporate math that just doesn't appear in 5e.

Personally, I would probably do a 5e beguiler as an unusual sorcerer origin, or (realistically) just use arcane trickster. But anyway, here are issues that I notice; things I don’t mention, I didn’t see anything wrong with, but that doesn’t mean there não são issues. I am still fairly new to 5e myself here.

Beguiler

• You have full access to all Illusion and Enchantment spells but the rest is limited, for these they must either enhance your own abilities, stealth, or be of some kind of utility for a skulk. (DM approval right now will make actual list when able)

  Nope, would not allow this. Writing a spell list is hard, but it is something that devo be done here. Otherwise you get weird interactions and unexpected power jumps. Nail this down before you expect to play the class.

• You know two 1st-level spells of your choice from the beguiler spell list. You learn an additional beguiler spell of your choice at each level except 12th, 14th, 16th, 18th, 19th, and 20th.

  To maintain the beguiler’s schtick, the spell list should be small, but the beguiler should know every spell on it, without need to prepare any subset of it. The actual number of spells on the list should be considered fairly carefully as a result: because the list is fixed and cannot be changed, it should be larger than the number of spells, say, a sorcerer learns, but only a bit more.

• These effects are only applied when the target has an enemy other than yourself within 5ft and that enemy is not incapacitated, or the target has disadvantage.

  This doesn’t really make sense in 5e. The first bit seems to be implementing a flanking rule, even though 5e doesn’t have one by default (and if the optional flanking rule is in use, it should be used here instead of this). More importantly, no one just “has” disadvantage, they have disadvantage on some particular roll. That isn’t a thing.

  I think the most sensible solution here is to take a cue from rogue:

  These effects are only applied when you would have advantage attacking the target.

Controlador de mentes

• At 1st level you gain Telepathy out to a range of 100ft to a single target at a time so long as they understand a language.

  This seems rather potent for 1st level. For context, 3.5e is generally (substantially) higher-power than 5e, and this wasn’t a 1st-level beguiler feature. Actually, it wasn’t a beguiler feature at all—it came from taking a level of the mindbender prestige class—but mindbender was so trivial for a beguiler to qualify for, and the beguiler gave up so little by taking that level, that many people fez think of it as a beguiler class feature—but a 6 nível class feature, not a 1st-level one.

Shadowcrafter

• At 1st level, you learn the minor illusion cantrip. In addition, it may create shadow material equipment and weapons.

  Unlike 3.5e, in 5e an extra cantrip is nothing to sneeze at. Granting a fairly-potent extra ability to that cantrip? Seems too good. Compare with the warlock’s Pact of the Blade: this is available 2 levels earlier, and is substantially better. Pact of the Blade is often considered underpowered, but ainda.

• Shadow Cloak

  This is a little unclear to me on how it actually works. Also, importantly, creatures gain advantage on their attacks against those who cannot see them (see all the warlocks using escuridão e a visão do diabo invocation)—if this is doing that, this is about as good as a spell and a 2nd-level class feature from the warlock; much too powerful.

• Shadow Illusion

  This just seems ridiculously potent and flexible to me. Shadowcraft mage was one of the most incredibly powerful classes in 3.5e, which as already noted, is higher-power than 5e.