Preface
What’s the point of writing a prompt guide for AI videos? Why not
dive straight into the tool?
That’s right: we learn generative AI through practice. However, the
cost of practice for AI video is an order of magnitude higher than
for testing image models or LLM tools, because video generation is
compute intensive, at least for now.
Free tiers of AI tools rarely show how well a video model really
works for your subject. Pay-per-second pricing adds up fast. Take
Veo 3 as an example: the full mode costs $6 for 8 seconds. By the
time you complete a test of 100 videos, that’s already a $600 bill.
There should be cost-effective ways to experiment with the
models before moving further into your creative use cases.
Here’s the great news: if you master the foundations of prompting
beforehand and identify areas with high success rates, you’ll
maximize the value of your tests and minimize waste. This will help
you plan your budget strategically and accurately for both your
personal explorations and commercial projects.
This prompt guide aims to boost your success rate with AI videos,
regardless of your subscription plan or API usage. Unlock your
imagination with AI video tools. Effective prompting is key.
This guide is based on the author’s hands-on experience, student
feedback, and theoretical knowledge gained from generating over
10,000 AI videos in a year.
As we are moving toward an agentic future, it’s even more
important to master a fundamental skill: communicating with AI
using your natural language.
© 2025 TYAI Pte. Ltd. All rights reserved. 3
1. GENAI VIDEO 101 [5-25]
2. KEY SUCCESS FACTORS [26-48]
3. CRAFTING EFFECTIVE PROMPTS[49-71]
4. KEYWORDS THAT MATTER [72-81]
5. MASTERING Image-to-Video [82-102]
6. HANDLING Video-to-Video [103-115]
7. STORYBOARD PROMPTING [116-136]
© 2025 TYAI Pte. Ltd. All rights reserved.
Case
Study
The story
behind this
viral video on
LinkedIn
© 2025 TYAI Pte. Ltd. All rights reserved. 21
Leonardo podcast:
How it’s made (1)
The Da Vinci podcast video, which became my most
successful LinkedIn post this year, was built on three key
elements: Story, Images, and Videos.
Story: The genesis of the idea was a museum visit years ago,
where I saw an exhibition of Leonardo da Vinci's
manuscripts. In his notes, I came across a device that looked
remarkably like a helicopter, which he called the "aerial
screw." According to Da Vinci, this device was inspired by the
water screw; if spun fast enough, it would lift off, just like a
modern helicopter. Thanks to modern AI, I no longer need to
brainstorm from scratch. Instead, I just need to mine my own
memory, my past photos and notes—for powerful stories
waiting to be told.
© 2025 TYAI Pte. Ltd. All rights reserved. 22
Leonardo podcast:
How it’s made (2)
Images: Generating a single AI video is simple, but creating a
coherent narrative requires high-quality inputs. This means
not just the text prompts, but also the reference images.
Using the Flux Kontext image model within an AI workflow
tool called Weavy, I generated multiple images for each
scene I envisioned: my avatar sitting with Leonardo da Vinci,
the aerial screw superimposed on a modern helicopter, and
several shots of Leonardo and me conversing inside and
outside the aircraft.
Character generation and iteration on Weavy
© 2025 TYAI Pte. Ltd. All rights reserved. 23
Leonardo podcast:
How it’s made (3)
Videos: The next step was bringing these static images to life.
I chose Google's Veo 3 for this task, as it excels at generating
naturalistic dialogue and motion for characters. It isn't
perfect, but it works exceptionally well. By the time you read
this, there may be even better models, but as of this writing
(July 2025), Veo 3 is the top contender. After generating the
clips on Google's Flow platform (using the cost-effective
'Fast' mode), it was a simple matter of using a standard
video editor to assemble them and add the final touches.
And that’s how it was done.
While there are AI agents and automated systems that can
handle most of these tasks for you, I believe in the power of
hands-on creation. As a practitioner, I see immense benefit
in actively engaging with every step of the creative process.
When you chat with the AI, direct the movements, and refine
the prompts, you are stepping into the role of a film director.
The more involved you are, the more the final output will
reflect your unique vision and authenticity.
Video generation and iteration on Flow
© 2025 TYAI Pte. Ltd. All rights reserved. 24
Get the full ebook here
tianyuxu.gumroad.com/l/videopromptbook
