2n

d
ed
iti
on
DALL-E 3
Full Playbook
A complete journey of styles, examples and tools

Tianyu Xu
Generative AI Explorer
April 2024 linkedin.com/tianyuxu
New in this edition
(Under ‘Part 1 Getting Started’)
1) Inpainting with DALL-E on ChatGPT
2) Style guides on OpenAI’s DALL-E GPT
3) Minor fixes and improvements
The updates will significantly enhance the
usability of the DALL-E images you create
and make your process more efficient.

Scan here for our Udemy

bestseller DALL-E course
This playbook is
dedicated to anyone who
1) wants to learn DALL-E 3
2) is keen to master image prompts
3) intends to generate unique images
4) seeks to explore multiple use cases
5) wishes to create images as they imagine
 PART 1
Getting Started
You can use DALL-E 3 for free
on Microsoft Copilot & Designer
Designer (via Bing) Designer (via Copilot)
bing.com/images/create copilot.microsoft.com

Designer (home page) Copilot Sidebar

designer.microsoft.com Copilot sidebar on Edge Browser

6
On mobile, you can use DALL-E 3
on apps like Bing, Copilot & Edge
Bing App Copilot App

7
Use DALL-E 3 on ChatGPT Plus to
get more help or customization
Chat with ChatGPT directly
Start with “create an image”

Use the Custom GPTs

Play with the top GPTs under ‘DALL-E’ category in the GPT Store

8
Explore style references on
OpenAI’s official DALL-E GPT
Locate this GPT in the GPT Store

9
Explore style references on
OpenAI’s official DALL-E GPT
Explore and select the style you need

Select the aspect ratio on the right

10
Explore style references on
OpenAI’s official DALL-E GPT
The output

Note: For advanced use cases, it’s

best to build your own Custom
Instructions and Custom GPTs.

Scan here for our Udemy

bestseller DALL-E course

11
Learn the basics of DALL-E
prompts from Copilot & Designer
Learn from the home page of Image Creator
bing.com/images/create

Follow the prompt templates on Designer

designer.microsoft.com

12
Start simple: try it out
Prompt: [angle/shot] photo of [subject]
eye-level high angle

macro shot wide angle

13
Start simple: try it out
Prompt: [subject] in the style of [style]
Starry Night 3D pixel

quilling cartoony 3D
14
Start simple: try it out
Prompt: sticker + [subject] + [text]
‘excuse me’ ‘what!’

‘really?’ ‘this is awesome!’

15
How about modifying the
DALL-E images directly?
It’s possible now with inpainting on ChatGPT:
you can modify specific parts of an image by
using selection tools and prompts.

Adjust brush Selection

Text prompt

17
Inpainting: practical use cases
#1 Change the backgound

18
Inpainting: practical use cases
#2 Change the character

19
Inpainting: practical use cases
#3 Refine the details (expressions)

20
Inpainting: practical use cases
#3 Refine the details (text and others)

21
Inpainting works on mobile, too

First, visit a DALL-E image on ChatGPT app

Selection tool

22
Inpainting works on mobile, too

Select an area, enter a prompt, and voilà ...

23
Inpainting: other features

Blend two images

You can upload a new image to the thread and ask
ChatGPT to combine the two images.
Note: it’s based on a combined prompt rather than editing the images directly.

24
Inpainting may not work with
all subjects and style blends
Selection: area on the right | Prompt: An alien taking selfie

Selection: area of the cat | Prompt: An alien in a spacesuit

25
Tips for beginners
Start with ‘low hanging fruits’
1) one human character in an image
2) popular art movements in history
3) pets, wild life, nature, architecture
4) general landscapes and cityscapes
5) common photo compositions & angles

then build on your prompts from there.

 PART 2
Photography
Style References
 What are the keywords (variables)
that affect DALL-E’s output on photos?

2.1 Angle/Shot

2.2 Style/Effect

2.3 Time of Day

2.4 Camera Film

2.5 Time/Era

2.6 Camera Brand

All images here were generated

using the basic prompt structures:
(THE CONSTANT) (THE VARIABLE)
“[subject], [photo-related style]”
OR “photo of the [subject], [style]”
28
 2.1 Angle/Shot
eye-level high angle

low angle wide angle

29
 2.1 Angle/Shot
macro shot close-up shot

medium shot long shot

30
 2.1 Angle/Shot
bird’s eye view worm’s eye view

sideways aerial
31
 2.2 Style/Effect
selfie group selfie

POV fisheye lens

32
 2.2 Style/Effect
selective color over the shoulder

frame in a frame group portrait

33
 2.2 Style/Effect
bokeh long exposure

neon lights steel wool

34
 2.2 Style/Effect
moonlit star trail

silhouette reflections
35
 2.3 Time of Day
sunrise morning

noon afternoon
36
 2.3 Time of Day
golden hour blue hour

night midnight
37
 2.4 Camera Film
Kodak Portra Kodak Tri-X

Kodak Ektar 100 Kodak T-MAX

38
 2.4 Camera Film
Fujifilm Pro 400H Lomo X-Pro

CineStill 800T LomoChrome Purple

39
 2.5 Time/Era
1890s Kodak No.1 1900s Brownie

1920s Rolleiflex 1930s Leica I

40
 2.5 Time/Era
1960s Nikon F 1970s Polaroid SX-70

1980s Canon AE-1 1980s Minolta Maxxum

41
 2.6 Camera Brand
Fujifilm X Canon EOS R

Sony A7 Nikon Z
Note: Camera brand does not seem to have
significant influence on the image output
2.6 (smartphone)
Camera Brand
iPhone Samsung

Huawei OPPO
Note: Camera brand does not seem to have
significant influence on the image output
 PART 3
Art Movement
Style References
 How do different keywords affect
DALL-E’s output on paintings?
Style Artist Artwork Era

3.1 Renaissance

3.2 Dutch Golden Age

3.3 Neoclassicism

3.4 Romanticism

3.5 Realism

3.6 Impressionism

3.7 Post-Impressionism

3.8 Song Dynasty

3.9 Ukiyo-e

All images here were generated

using the basic prompt structures
(THE CONSTANT) (THE VARIABLE)
“[subject] + [style/artist/artwork/era]”
45
 3.1 Renaissance
in the style of by Leonardo Da Vinci
Renaissance

in the style of Early 16th century

‘Mona Lisa’ Italian painting
46
 3.2 Dutch Golden Age
in the style of by Johannes Vermeer
Dutch Golden Age

in the style of 17th century

‘Girl with a Pearl Earring’ Dutch painting
47
 3.3 Neoclassicism
in the style of by Jacques-Louis David
Neoclassicism

in the style of ‘Napoleon Late 18th century

Crossing the Alps’ French painting
48
 3.4 Romanticism
in the style of by Eugène Delacroix
Romanticism

in the style of ‘La Liberté Early 19th century

guidant le peuple’ French painting
49
 3.5 Realism
in the style of by Gustave Courbet
Realism

in the style of Mid 19th century

‘Le Désespéré’ French painting
50
 3.6 Impressionism
in the style of by Claude Monet
Impressionism

in the style of 1870s French painting

‘Impression, Sunrise’
51
3.7 Post-Impressionism
in the style of by Vincent Van Gogh
Post-Impressionism

in the style of 1890s French painting

‘The Starry Night’
52
 3.8 Song Dynasty
in the style of Song by Emperor Huizong
Dynasty painting

in the style of 12th century

‘Auspicious Cranes’ Chinese art
53
 3.9 Ukiyo-e
in the style of Ukiyo-e by Katsushika Hokusai

in the style of ‘The Great 1830s Japanese art

Wave off Kanagawa’
54
 So, which variables influence
the output of the AI painting?

All of them do.

Depending on the artwork's style and
the artist's popularity, the degree of
adherence to that style will vary.
 PART 4
Art Medium
Style References
 What are the art mediums (variables)
that affect DALL-E’s output on images?

4.1 Textile

4.2 Sculpture

4.3 Printmaking

4.4 Digital Art

4.5 Drawing

4.6 Painting

4.7 Other

All images here were generated

using the basic prompt structure:
(THE CONSTANT) (THE VARIABLE)
“[subject] + [art medium]”

57
 4.1 Textile
batik quilting

lace making feltmaking

58
 4.1 Textile
tapestry weaving embroidery

knitting macrame
59
 4.2 Sculpture
clay gold

bronze casting stone calving

60
 4.2 Sculpture
wood carving sand

ice sculpting glass blowing

61
 4.3 Printmaking
etching print woodcut print

screen printing linocut printing

62
 4.3 Printmaking
monoprinting collagraphy

drypoint print mezzotint print

63
 4.4 Digital Art
CGI concept art

algorithmic art digital collage

64
 4.4 Digital Art
digital painting vector graphics

pixel art 3D modeling

65
 4.5 Drawing
soft pastel crayon

oil pastel marker

66
 4.5 Drawing
colored pencil charcoal

chalk graphite
67
 4.5 Drawing
silverpoint pencil

ballpoint pen ink

68
 4.6 Painting
tempera encaustic

spray acrylic
69
 4.6 Painting
oil painting ink wash

water color gouache

70
 4.6 Painting
casein fresco

colored ink brushwork

71
 4.7 Other
napkin drawing rice art

paper cutting origami

72
 4.7 Other
porcelain quilling

graffiti texture art

73
 PART 5
Visual Storytelling
Style References
What are the visual storytelling keywords
that affect DALL-E’s output on images?

5.1 Category

5.2 Art Medium

5.3 Drawing Style

5.4 3D Style

5.5 Theme

5.6 Era Style

All images here were generated

using the basic prompt structure:
(THE CONSTANT) (THE VARIABLE)
“[subject] + [style]”

75
 5.1 Category
manga animation

storyboard comic
76
 5.1 Category
cartoon webtoon

anime illustration
77
 5.2 Art Medium
cartoon, ink drawing cartoon, pencil drawing

cartoon, marker cartoon, watercolor

drawing
78
 5.2 Art Medium
cartoon, crayon cartoon, gouache

cartoon, digital media cartoon, chalk

79
 5.3 Drawing Style
old cartoon modern cartoon

classic manga modern anime

80
 5.3 Drawing Style
chibi drawing minimalistic cartoon

caricature line drawing cartoon

81
 5.3 Drawing Style
manga, kodomo manga, shonen

manga, shojo manga, mecha

82
 5.4 3D Style
CGI realistic CGI

abstract CGI stylized CGI

83
 5.4 3D Style
3D anime 3D pixel

cartoony 3D low-poly 3D
84
 5.4 3D Style
clay animation wooden puppet

horror 3D retro futurism 3D

85
 5.5 Theme
cartoon, cyberpunk cartoon, steampunk

cartoon, noir cartoon, post-

apocalyptic
86
 5.5 Theme
cartoon, fantacy cartoon, sci-fi

cartoon, comedy cartoon, horror

87
 5.5 Theme
comic, fantasy comic, sci-fi

comic, cyberpunk comic, steampunk

88
 5.5 Theme
comic, noir comic, mystery

comic, romance comic, post-

apocalyptic
89
 5.5 Theme
manga, cyberpunk manga, steampunk

manga, noir manga, post-

apocalyptic
90
 5.5 Theme
manga, fantasy manga, sci-fi

manga, adventure manga, horror

91
 5.6 Era Style
cartoon, 1920s cartoon, 1940s

cartoon, 1990s cartoon, 2000s

92
 5.6 Era Style
comic, 1920s comic, 1950s

comic, 1990s comic, 2000s

93
 PART 6
Design & Text
Style References
What are the design-related themes and
keywords that work well with DALL-E?

6.1 Stickers

6.2 Logos

6.3 App Icons

6.4 Emojis

6.5 Patterns

6.6 Word Art

6.7 White Boards

6.8 Clean Backdrop

6.9 Fashion

95
 6.1 Stickers
Basic prompt: sticker + [subject] + [text]
‘excuse me’ ‘what!’

‘really?’ ‘this is awesome!’

96
 6.2 Logos
Basic prompt: [style] + logo + [subject]
hand-drawn geometric

dynamic gradient
97
 6.3 App Icons
Basic prompt: [style] app icon + [subject]
minimalism skeuomorphism

flat design 3d graphics

98
 6.4 Emojis
Basic prompt: [theme] emoji pack + [subject]
facial expressions business theme

crying professions
99
 6.5 Patterns
Basic prompt: [pattern name] pattern + [subject]
floral abstract

paisley stripes
100
 6.5 Patterns
Basic prompt: [pattern name] pattern + [subject]
damask tartan

batik argyle
101
 6.6 Word Art
Basic prompt: [subject] + ‘sale’ + [style]
neon paper cut

ice and frost fire and flame

102
 6.7 White Boards
Basic prompt: [subject] + holding a white [object]
board business card

monitor screen smartphone screen

103
 6.8 Clean Backdrop
Basic prompt: [design] + white background
3-seater sofa children’s bed

bamboo wardrobe dining table set

104
 6.8 Clean Backdrop
Basic prompt: [design] + dark background
3-seater sofa children’s bed

bamboo wardrobe dining table set

105
 6.9 Fashion
Basic prompt: [subject] + [type of clothes]
jacket t-shirt

hoodie dress
106
 6.9 Fashion
Basic prompt: [subject] in a dress + [material]
cotton wool

silk recycled materials

107
 6.9 Fashion
Basic prompt: [subject] in a dress + [style/pattern]
geometric chevron

houndstooth stripes
108
 PART 7
Capturing Actions
 What are the actions that
work well with DALL-E?

7.1 Posture

7.2 Expression

7.3 Movement

7.4 Interaction

7.5 Gesture

All images here were generated

using the basic prompt structure:
[character] + [action with context]
(THE CONSTANT) (THE VARIABLE)

110
 7.1 Posture
perching cautiously lunging aggressively
on a fence at a rival

balancing precariously kneeling respectfully

on a narrow ledge before a statue
111
 7.1 Posture
sitting regally sprawling lazily
on a throne on a couch

leaning curiously cowering timidly

against a bookshelf under a table
112
 7.2 Expression
smiling gleefully blushing shyly after
with a ball of yarn receiving a compliment

scowling at grinning mischievously

a barking dog with a feather in mouth
113
 7.2 Expression
sneering disdainfully frowning at
at a rival cat a broken vase

yawning sleepily appearing surprised

in a warm sunbeam by a loud noise
114
 7.3 Movement
leaping playfully skipping joyfully
after a mouse through a flower field

striding confidently pacing nervously

across a field before a challenge
115
 7.3 Movement
falling clumsily crawling cautiously
from a tree through a tunnel

dancing playfully stretching luxuriously

to music in a sunbeam
116
 7.4 Interaction
fighting playfully chasing a butterfly
with an orange cat through a meadow

nuzzling affectionately tossing a ball

against its owner's leg in the air
117
 7.4 Interaction
lifting a riding on the back
collapsed wall of a robotic horse

sharing food watching a

with a robotic dog holographic projection
118
 7.5 Gesture
pointing excitedly waving enthusiastically
at a bird in a tree to a friend

shrugging with signaling a secret

indifference with a wink
119
 7.5 Gesture
holding a precious clapping happily
object carefully at a performance

scratching its head tapping its chin

in confusion thoughtfully
120
 PART 8
Leveraging Tools
You can add more context
to a prompt with ChatGPT
Prompt for ChatGPT: create an image of the [subject] in the
style of the [name of the painting] in the same context with
descriptive keywords
BEFORE
(without using ChatGPT)

painting of a pure white cat with

blue eyes dressed in yellow, in the
style of 'The Great Wave off
Kanagawa'.

Ukiyo-e style painting illustrating a

pristine white cat with striking blue
eyes, dressed in a yellow garment,
seated on a boat amidst the
powerful waves reminiscent of 'The
Great Wave off Kanagawa'. The
composition captures the
movement of the sea with
woodblock print textures and
includes the iconic Mt. Fuji in the
backdrop, all in harmony with the
traditional Japanese art style.
AFTER
122
You can add more context
to a prompt with ChatGPT
Prompt for ChatGPT: create an image of the [subject] in the
style of the [name of the painting] in the same context with
descriptive keywords
BEFORE
(without using ChatGPT)

“painting of a pure white cat with

blue eyes dressed in yellow in the
style of 'Napoleon Crossing the
Alps'.”

“A pure white cat with blue eyes

dressed in a yellow Napoleonic
uniform, posed heroically atop a
rearing horse amidst a snowy
mountain backdrop. The scene is
reminiscent of Jacques-Louis
David's 'Napoleon Crossing the
Alps', with dramatic lighting and a
dynamic composition that conveys
the sense of movement and power
associated with the famous
painting.”
AFTER
123
You can also turn the images
into designs with Designer
Step 1: Start
with ‘Design
Creator’

Step 2: Choose
‘Generate image’
to use DALL-E 3

Note: Designer is currently at Preview stage and not commercial ready.

Reference: https://designer.microsoft.com/termsOfUse.pdf

124
You can also turn the images
into designs with Designer
Step 3: Describe the design or ‘Try a prompt’

Step 4: Click ‘Customize design’ to continue editing

125
Or make further edits to your
DALL-E generated image
Try “Remove background” on Designer

126
Or create a full storyboard with
DALL-E for generative AI videos
Generate an image with DALL-E
“a white cat driving a red car in the desert”

Upload the image to an image-to-video tool

Runway Gen2 Pika 1.0

127
PART 9
Scaling with
Custom GPTs
 Custom GPTs with DALL-E are
powerful tools for storytelling
and generating images at scale
I developed a series of Custom GPTs,
each designed as a tuned version of
DALL-E to fulfill specific roles.

9.1 9.2 9.3

9.4 9.5 9.6

How to build a Custom GPT?
Method 1: Chat with GPT Builder bot

Method 2: Configure the GPT directly with instructions

130
 9.1 Portrait Paws
https://chat.openai.com/g/g-W9TOAB2Xg-portrait-paws

Why should you try it?

Convert a photo of your pet or
provide a description to create
portraits in popular art styles
Demo: Various styles based on uploaded image
UPLOADED IMAGE
 9.2 Manga Meowster
https://chat.openai.com/g/g-D4iz6JTzp-manga-meowster

Why should you try it?

Create your own story and
visualize it using consistent
character and style.
Demo: Story of a little frog looking for mama
 9.3 Trip Meowster
https://chat.openai.com/g/g-5wVHKO7cK-trip-meowster

Why should you try it?

Imagine your family trip and
receive a tailored itinerary
with up-to-date infomation.

Demo: 6-day family trip to Seoul from Singapore

 9.4 Meowart History
https://chat.openai.com/g/g-m4ZGswnYG-meowart-history

Why should you try it?

Discover world art history and
reimagine artworks with cat
master in a meowseum
Demo: Ancient sculptures around the world
 9.5 Crayon Explainer
https://chat.openai.com/g/g-Qa1kcQkZg-crayon-explainer

Why should you try it?

Explain and draw everything
from rocket science to life
science for a 5-year-old
Demo: Life, rocket, REITs, LLM, photosynthesis
 9.6 Purrfect Pages
https://chat.openai.com/g/g-D7WG0zVdD-purrfect-pages

Why should you try it?

Convert any subject into
children’s coloring pages: just
say the subject in the chat.

Demo: Park Güell, Singapore CBD, various animals

Remember to have fun
during the trying and
building processes!
Resources
Usage

https://openai.com/pricing
Pricing (API)

Official guides and usage policies

https://help.openai.com/en/articles/9055440-editing-your-images-with-dall-e
https://help.openai.com/en/articles/6516417-dall-e-editor-guide
https://help.openai.com/en/collections/3643409-dall-e-content-policy
https://help.openai.com/en/articles/6468065-dall-e-content-policy-faq
https://openai.com/policies/usage-policies
https://www.bing.com/images/create/contentpolicy
https://designer.microsoft.com/termsOfUse.pdf

138
Keen to master DALL-E?

Scan here

More GenAI playbooks

Tianyu Xu
Generative AI Explorer
linkedin.com/tianyuxu