I went in with high hopes for this book ($15 for homebrew content with only a couple reviews, but neat ideas) and was hoping the authors would "wow" me with good mechanics to match some of the cool ideas put forth in the description (was especially excited about the Diabolist, Feywalker, Noble). Unfortunately, I can not recommend this to anyone looking to expand their table's class selection.
An overarching issue is that most of the classes thematically (and some, mechanically) appear as though they easily could have been (and due to a variety of reoccuring balance issues, perhaps should have been) created within the framework of existing classes. Diabolist stands out especially in this regard for both, the class description presents no real difference between it and the Warlock and the core mechanics (Necro debuffs, summons) could easily have been a Patron and/or Pact Boon combination with these simply giving access to summoning spells/abilities and a Debuff effect for certain spells (with the addition of some Invocations, if that would be easier to streamline). I simply do not understand some of the mechanical choices made to represent some of the ideas, which themselves take unnecessary steps to distance themselves from existing classes (and sometimes in flat out poor ways, such as elevating the class in its description text above existing classes).
There's is a correlation within this document that the more an ability deviates from existing framework/class abilities the more mechanical and balance issues it tends to have. Some abilities (The Noble's Expertise comes to mind, which grants expertise to 3 skills rather than 2 like every other Expertise ability in the 5e) are blatantly overpowered compared against their counterparts. The Cardcaster (as fond as I may be of its inspiration) is perhaps the most unique class produced in the document and, unsurprisingly, suffers the most in terms of balance, unnecessary complexity (read: fat could have been trimmed to produce the same desired effect) and poor design. This correlation is entirely consistent throughout the document.
While not as critical as the mechanics, the flavor text is often poorly written in that it references mechanical concepts rather than taking the opportunity to highlight what such a character of this class might be like, might believe. Referring to "lawful slavers" and "chaotic" characters does not accomplish this at all. Combined with how the classes fail to differentiate themselves effectively from their core class counterparts, I get the feeling the concepts did not successfully transition from bundles of custom-tailored mechanics to fully realized and rich classes: skill kits from which different sorts of believably rounded people might grow and develop. That might not bother you if you're looking for a new class to tinker with, but for some it is a pretty big let down.
I did find the Alchemist class to have some of the more creative ideas but it too suffers from horrible balance issues that should have been very apparent during playtesting (scaling at 1st, 5th, 9th, 13th, 17th levels instead of 1st, 5th, 11th, 17th like every other spell effect in the game? And throwing multiple of them as well?). I also thought the formatting and art was very well done and the quality for these was consistent through the document. Unfortunately, the design problems were too great to ignore. That's perhaps the best way to describe this document as a whole: "Poorly Executed Design".
I am a huge fan and supporter of homebrew developers and love seeing the community produce new and exciting content to try at my table, but the problems here are simply far too much for a normal 5e table to incorporate without immense revision which might as well be an overhaul. When you charge professional prices ($15 for 70~ pages of content) you need to take the time to ensure your work can at least approach existing professional standards in the field. This does not even begin to approach that qualifying level of quality.
I would recommend you spend your money elsewhere.
[2 of 5 Stars!]