Ritzhard's Quick Guide to Yodayo
Who is this for?
- Anyone that looked at my pictures on JanitorAI and wants to know how I did it.
- Anyone that is interested in generating their own images.
- Just anyone in general.
Why I made this
- Generating AI images might seem difficult, but in reality, it's quite easy.
- I've been asked a few times about how I generate my images, so here we are.
What this is — and what it is not
This is:
- A casual guide based on personal experience with Yodayo and generating images in general.
- A small collection of what I've learned so far.
- A quick guide so that you can start generating.
This is not:
- Written by an expert, nor will it guarantee you perfect results.
- A guarantee that following this guide will lead you to creating perfect images that fit your vision.
- A step by step where I explain every tag/prompt, that'll take too long xdd.
Most of what I know came from lots of trial, error, and testing.
Take what helps you, skip what doesn’t, and develop your own workflow.
First of all, what is Yodayo?
- Yodayo originally started as a VTuber community.
- In its early days, you'd actually see Nakiri Ayame greeting you on the front page.
- It evolved into an AI image generation site around early 2023.
There's Yodayo and Moescape, so what is what?
Yodayo and Moescape are the same platform under different names:
- Yodayo: Allows NSFW images
- Moescape: SFW-only version of the same platform
How does the currency system work?
When you first make an account on Yodayo, here's what you start with and what you can expect daily:
- You begin with 100 Yobeans (the site currency for generating).
- As a normal user, you receive:
35 beans daily- 50 beans daily (as of 05/06/2025 — updated)
Note: Daily beans do not stack. If you don’t use them, they’re gone the next day.
How much does it cost to generate images?
- Each image costs between 2–5 Yobeans
- High-resolution (hi-res) generations can cost up to 10 Yobeans
So how many images can you realistically generate per day?
- With 50 beans daily, you can expect:
- Around 10–13 regular generations
- Fewer if you're generating hi-res images
Which is honestly not a lot.
Quick Overview of Yodayo UI
Everything you need to know when you first land on the image generation screen
Getting There
- Once you've created an account, go to the AI Art dropdown in the top bar
- Select AI Art to enter the main generation UI
You’ll see the following main components:
Prompt, Negative Prompt, Settings, Canvas Size, Hires.fix, Model Hub, Spells
Prompt
- This is where you type what you want the AI to generate
- Example:
"a big tiddy goth gf"
- And yes, that will generate something decent
- But if you want better results, you’ll need more than just basic prompts
Negative Prompt
- Input what you don’t want the AI to generate
- Useful for avoiding things like:
- Extra limbs
- Bad hands
- Bad anatomy
- Blurry eyes
Settings
This section controls how the image is generated.
You really only need to understand:
- Steps:
- Number of steps the model takes before finalizing the image
- More steps ≠ better
- 35 is the sweet spot for me
- More than 35 may cause issues: extra fingers, multiple pupils, artifacting
- Scale:
- Controls how literally the prompt is followed
- Lower scale = the model takes artistic liberty
- Higher scale = sticks closer to your prompt
- Example:
- High scale (12.5): More literal output
- Low scale (6.5): More interpretive result
- Seed:
- Determines the randomness
- Using the same seed gives you the exact same image every time
- Change it to experiment or re-roll
Canvas Size
- Controls the resolution / dimensions of the generated image
- Self-explanatory
Hires.fix
This is for upscaling and improving image quality after you get a good result.
- Upscale:
- Increases image dimensions
- Hires Steps:
- Adds pixel detail (Similar-ish to the steps earlier)
- Denoising Strength:
- Controls how much the model changes from the base image
- Lower = less deviation, just pixel boost
- Higher = more edits to the image
I mostly use this when I want to turn a result I really like into a final, hi-res version.
Model Hub
- Lets you pick which model you want to use
- Models affect the style, anatomy, detail, etc.
My favorite: [Illustrious] Nova Orange XL
- You can experiment to find one that suits your taste
Spells
- They are modifiers that inject specific traits or styles into the image
- These affect clothing, anatomy, composition, etc.
If you like how my images look, it’s because I use Ratatatat74-style here
Enough yapping about the site... let's start prompting.
So, how do you actually start generating images on Yodayo?
Let’s say I want to make an image of a goth girl with big tiddies, using this image as inspiration — with a mask on.
Step 1: Look at the reference image
I usually take a look at the danbooru tags. Here's the link to that image with all the tags: Baddie Miku
Step 2: Write the Prompt
Here’s the exact prompt I used for this look:
newest, highres, absurdres, (masterpiece), best quality, highly detailed, high quality, expressive eyes, Ratatatat74, Enchanting Eyes,
BREAK perfect face, intricate face, beautiful, (black eyes, red pupils), eyelashes, (round eyewear, black eyewear), (smirking, open mouth, sharp teeth, tongue out), (hair over eye, long hair, black hair, red streaks, multicolored hair, bob cut), (asymmetrical hair, long bangs, messy bangs, asymmetrical bangs), (eyebrow cut, thick eyebrows), (black lips), (spiked choker, choker, black choker), (earrings, cross earrings, ear piercings), (mouth mask, mask pull),
BREAK (leather tanktop, harness, chest harness, o-ring, o-ring harness, strap between breasts), (black sleeves, fishnet sleeves), black wrist cuffs, (black shorts, shorts, black panties, high leg panties), (fishnet pantyhose, single leg pantyhose, goth fashion), (fur-trimmed jacket, long jacket, oversized jacket),
BREAK (from front, from above), (night time, outdoors, street), (solo focus, solo female, 1girl, female, close-up), medium breasts, thighs, stomach, navel, tall female, hand on own hip, hand on own chest, leaning forward,
Step 3: Add the Negative Prompt
This tells the model what you don’t want. This is my general settings, but feel free to add in things you wouldn't want to see in the image (nails, other hair colors, extra details on weapons, etc)
modern, recent, old, oldest, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, very displeasing, (worst quality, bad quality:1.2), sketch, jpeg artifacts, signature, watermark, username, simple background, conjoined, bad AI-generated, deformed anatomy, bad fingers, more than 5 fingers in one hand, less than 5 fingers in one hand, fat, chibi, chubby, blurry, extra hands, muscular girls, (bandage:1.1), tape, (low quality:1.4), monochrome, trademark, multiple view, EasyNegative, badhandv4, (solid fabric:1.3), (opaque clothing:1.3), (thick fabric:1.2), (non-transparent:1.3), (not see-through:1.3), (cotton texture:1.2), (normal shirt:1.2), ((Inscriptions)),
Step 4: Choosing your model.
My Model is [Illustrious] Nova Orange XL.
Step 5: Choosing your Settings
My settings are:
Step 6: Choosing your Loras/Spells:
My Loras/Spells are:
And this should be the result, and this was how I generated Cassandra:
Step 7 (Optional): Choose whether or not you want to increase the resolution of the image:
This is the hires settings I used:
And this is her after Hi-res:
What does BREAK actually do?
Before I dive into what each part of the prompt does, I'll quickly talk about BREAK Prompts.
What is BREAK?
In the simplest terms:
BREAK is like starting a new paragraph in an essay.
Every prompt chunk you write after a BREAK is treated like its own group or section. This helps reduce:
- Prompt bleeding
- Prompt artifacts
- Confused or overlapping features
Think of () like “ending a sentence” and BREAK as “starting a new paragraph.”
Here’s how I organize it:
newest, highres, absurdres, (masterpiece), best quality, highly detailed, high quality, Ratatatat74, Enchanting Eyes,
This is where I usually put general quality tags and LoRAs that set a style (e.g., Ratatatat74, Nyantcha, Raita) or apply general filters (like Enchanting Eyes, EPS, etc.)
BREAK perfect face, intricate face, beautiful, expressive eyes, (black eyes, red pupils), eyelashes, (round eyewear, black eyewear), (smirking, open mouth, sharp teeth, tongue out), (hair over eye, long hair, black hair, red streaks, multicolored hair, bob cut), (asymmetrical hair, long bangs, messy bangs, asymmetrical bangs), (eyebrow cut, thick eyebrows), (black lips), (spiked choker, choker, black choker), (earrings, cross earrings, ear piercings), (mouth mask, mask pull),
I usually group everything shoulders and up in one break prompt so that it's both easier for me to track and for the the model to generate. My preferred order is Face details > Hair> Makeup > Accessories.
BREAK (leather tanktop, harness, chest harness, o-ring, o-ring harness, strap between breasts), (black sleeves, fishnet sleeves), black wrist cuffs, (black shorts, shorts, black panties, high leg panties), (fishnet pantyhose, single leg pantyhose, goth fashion), (fur-trimmed jacket, long jacket, oversized jacket),
Here’s where I usually group clothing, and usually I go top down, so upper body, then lower body.
BREAK (from front, from above), (night time, outdoors, street), (solo focus, solo female, 1girl, female, close-up), medium breasts, thighs, stomach, navel, tall female, hand on own hip, hand on own chest, leaning forward,
Final part of the prompt.
This controls:
- Angle and scene (e.g., from front, night time)
- Subject placement (1girl, solo focus)
- Body pose or emphasis on including certain body parts (stomach, navel, leaning forward)
Including things like stomach/navel helps hint that the clothes should be cropped or revealing.
Why Yodayo? Why not Pixai? Tensor? CivitAI? Or even local generation?
Good question!
As one of the older sites, Yodayo definitely has its flaws — and not many pros.
But the few pros it does have are why I still use it today.
It's not the best site out there, and in terms of raw power, here's my personal ranking:
Locally generated > CivitAI > Tensor > Pixai > Yodayo
Pros:
- No LoRA cap: Whether you use 1 LoRA or 100, it costs the same amount of beans. No penalty.
- Understands BREAK prompts better than quite a lot of others.
- Their prompting style is very close to danbooru style of tagging, so it's a lot more straightforward.
Cons:
- 50 beans per day: Not generous at all — and no quick way to earn more as a free user.
- Understands BREAK prompts better than quite a lot of others.: One mistake in prompt and it can change how the image look xdd. You need to be quite thorough.
That should be everything!
Of course, there’s way more I could talk about — but this is meant to be a quick guide, and I think it’s already long enough.
If you have any questions, feel free to hit me up on Discord: @mepba