Verified purchaser
I really wanted to like this
I tried a free ebook. The result was bad. Then I bought tier-2.
I uploaded articles I've written over the years to create my own model. It took a long time.
Then I carefully crafted a Table of Contents and a prompt using Gemini, Claude and ChatGPT with back and forth improvements. My prompt was rated as 95% good. 24 hours or so later, I received an awful lot of unusable stuff. To be able to publish that would require a lot of work and the result would still be bad.
I have several prompts that I use with AI engines, that can create a full 2500 words chapter that barely needs editing. This means that I can produce a good quality 50K words book in less than 5 days, ready to publish. In contrast, editing the book from youbooks would take me more than 5 days and the result would be far from good.
But I tried again with another book. This time I spent less time on the prompt, copy-pasting almost straight from Claude. The result was underwhelming, if slightly better. Style problems, poor narration, poor structure - it was an eyesore.
So I tried another one. This time I opted for "expensive models" since the not expensive ones are so bad. In his answers to the questions, the product's owner recommends avoiding this expensive 4x model altogether. The result was much better than the two previous tries. This time I got something that still needs heavy editing, but probably not more than 2 days of work. Is it good ? Not really, but it's not bad either.
What is the future for this tool ? I'm not sure. I need to give it another couple tries but credits renew only on the 26th and I bought the product on the 27th so I'm gonna have to wait a full month before I can do additional checks. For now I will give it two tacos and I'll change my rating if I get better results.
You should buy it only if you are prepared to edit heavily. As a former journalist, editor-in-chief and owner of a content agency, I know how to edit long form documents.
Or if you are not interested in editing, just in publishing mounds of copy.
The question is : would I get a better book if I use my own prompts with Gemini or Claude ? If that's the case then I have no use for Youbooks. Create a prompt with a TOC. Upload it to Gemini with instructions to write one chapter after the other while keeping the whole thing coherent. It's not rocket science and you get to check each chapter after it's created, which gives you ample time to correct things for the following chapters. Well, I'll perform additional checks in a month...
Ioannis_Youbooks
Edited Oct 31, 2025Perhaps the best bad review we have received :)
You are right that Youbooks will not produce edit-ready copy. ChatGpt does not produce edit ready articles. One-shotting a book is a 100x harder task. This is not an excuse. It's a property of the product that many positive reviews have also pointed out. If you care about quality and tailoring to your taste, you need to edit as you correctly pointed out.
However, you are not right to suggest that prompting Gemini to produce the next chapter and the next will produce a similar book. I am sure you are saying this in good faith. And I am sure you think it could work. And the matter of the fact is that there is a bunch of AI book maker apps that are doing just that.
Youbooks does not. I have been through enough evals (hundreds by now over a period of a year) to make me want to cry false at your statement. I understand you think this is the case and you have the best intentions. But you come from a place of well-intended assumptions vs a lot of trial and experience. Try this. Take a topic that is niche enough to have a chance of selling (not another book on how to make passive income). Write a book using your technique with Gemini. Do the same with youbooks (don't use human models, leave internet search ON). Produce 80K words of text that is an average nonfiction book. Ask a strong model like ChatGpt Pro to rate the youbooks output and the Gemini stitched output.
I appreciate your honest opinion. If it was only a matter of taste, I would be happy to also agree with the last point (what's gold for some is poop for others). It's only because we can run AI based evals to get some sense of an un-biased feedback, that I find the courage to push back.