prompt: you will be given a book chapter or article text your task is to select three to four key moments from the text or the main subject of an article and create visual descriptions of these scenes as image prompts for an AI image generator follow these steps : 1- read and analyze the following text you are given carefully take note of characters and location details. 2- step two select three to four key moments from the text these should be significant scenes or events that are crucial to the narrative or content or they could be a singular character portrait. 3- for each key moment create a visual description to be used as an image prompt follow these guidelines for each prompt: a- Begin with the most important part of the prompt (the main subject) near the beginning. b- focus on one to two subjects maximum describing each separately. c- include key details separated by commas The Prompt doesn't need to be grammatically correct it can be a list of image details. e- be specific and descriptive including details about setting character actions and atmosphere. f- make sure the characters and locations include accurate descriptive details from the text and that these are consistent across your three to four prompts. 4- format your output as follows: prompt 1: [your first image prompt here] prompt 2: [your second image prompt here] prompt 3: [your third image prompt here] 5- end the prompt with the following " colouring book, simple illustration, fancy, historical novel style" remember each prompt should be concise yet Vivid capturing the essence of the key moment in a way that can be visually represented, avoid abstract Concepts and focus on concrete visual elements.
@@Truth-reins-free Left-click and hold the mouse until you highlight the entire text, then right-click on it and choose copy. I assume you are a typewriter person?
This is awesome stuff. One thing to get the red streak is to RUN her character first, to get her as you wanted , then use the text prompts with her as an Image Ref, and Say, "A bright red Colored Highlight streak in her hair that goes from the scalp of her hair to the tips" you can also add Style Ref images to the prompt if you want a consist painting style, that will maintain the same look throughout your book.
Newbie on this particular stuff. How would I get the system to look to an image that I have on my computer? Do I need to upload to a cloud, and set a URL for the link?
Great idea. I did just that a few months ago for a book cover to my dads memoirs when he was growing up in Poland. I took a whole chapter, reduced it with chatgpt, plug it in as a prompt and I get an awesome book cover the first attempt! Im also using my dads voice to make an audio version with elevenlabs. A very long ways from being done but its exciting.
Great! I used a derivation in both Claude and Gemini and they worked really well. Gemini was able to send it right to image. One thing I added was a step where I pick which setting to add to end to designate the style of image. Based on picking from a list, the AI added it efficiently and generated the image. An improvement I would love to add is a list of all the various settings that could be used for the style. Also, a selection prompt for things like orientation and any other choices that one would make to the image generator (such as Midjourney). Any idea where such a list may already exist?
If you subscribe to a program like Flux or whatever, they have tools that allow you to do a face swap. Generate the image with the prompt, then swap the face with your character's.
@@Ank3rmanCan you do this for multiple characters? I've been looking for a way to create several characters that I'd like to turn into a mixed reality educational cartoon, where the characters interact with my real world but I'm not sure we are quite there yet. The characters would need to be consistent between episodes, and potentially overlap video or images in the real world.
@@tracy419 I believe so. Matt Wolfe has a recent video on AI he uses every day and he went through the process for creating a thumbnail in flux. You just paint over which face you want to swap and upload the reference image. It's definitely doable with today's technology but it might require just a little effort and know-how.
Here you go: You will be given a book chapter or article text. Your task is to select 3-4 key moments from the text (or the main subject of an article) and create visual descriptions of these scenes as image prompts for an AI image generator. Follow these steps: 1. Read and analyze the following text you are given carefully. Take note of character and location details. 2. Select 3-4 key moments from the text. These should be significant scenes or events that are crucial to the narrative or content. Or they could be a singular character portrait. 3. For each key moment, create a visual description to be used as an image prompt. Follow these guidelines for each prompt: a. Begin with the most important part of the prompt (the main subject) near the beginning. b. Focus on 1-2 subjects maximum, describing each separately. c. Include key details, separated by commas. d. The prompt doesn't need to be grammatically correct; it can be a list of image details. e. Be specific and descriptive, including details about setting, characters, actions, and atmosphere. f. Describe characters without naming them (e.g., "a young man standing in a field," not "Tom standing in a field"). g. Include the general time period or setting, especially if it is not modern (e.g., "medieval village," "Victorian street"). h. Make sure the characters and locations include accurate descriptive details from the text and that these are consistent across your 3-4 prompts. 4. Format your output as follows: Prompt 1: [Your first image prompt here] Prompt 2: [Your second image prompt here] Prompt 3: [Your third image prompt here] Prompt 4: [Your fourth image prompt here] 5. End each prompt with the following: "watercolor children's illustration, simple illustration, wonderous" Remember, each prompt should be concise yet vivid, capturing the essence of the key moment in a way that can be visually represented. Avoid abstract concepts and focus on concrete, visual elements.
Here it is: You will be given a book chapter or article text. Your task is to select 3-4 key moments from the text (or the main subject of an article) and create visual descriptions of these scenes as image prompts for an AI image generator. Follow these steps: 1. Read and analyze the following text you are given carefully. Take note of character and location details. 2. Select 3-4 key moments from the text. These should be significant scenes or events that are crucial to the narrative or content. Or they could be a singular character portrait. 3. For each key moment, create a visual description to be used as an image prompt. Follow these guidelines for each prompt: a. Begin with the most important part of the prompt (the main subject) near the beginning. b. Focus on 1-2 subjects maximum, describing each separately. c. Include key details, separated by commas. d. The prompt doesn't need to be grammatically correct; it can be a list of image details. e. Be specific and descriptive, including details about setting, characters, actions, and atmosphere. f. Describe characters without naming them (e.g., "a young man standing in a field," not "Tom standing in a field"). g. Include the general time period or setting, especially if it is not modern (e.g., "medieval village," "Victorian street"). h. Make sure the characters and locations include accurate descriptive details from the text and that these are consistent across your 3-4 prompts. 4. Format your output as follows: Prompt 1: [Your first image prompt here] Prompt 2: [Your second image prompt here] Prompt 3: [Your third image prompt here] Prompt 4: [Your fourth image prompt here] 5. End each prompt with the following: "watercolor children's illustration, simple illustration, wonderous" Remember, each prompt should be concise yet vivid, capturing the essence of the key moment in a way that can be visually represented. Avoid abstract concepts and focus on concrete, visual elements.
What if you run a local generator like ComfyUI for example? What model would be good to use to "try" and generate a novel style images from CivitAI? I have a ton of models so far that does SFW and NSFW of course. But, I'm willing to get a model that can do the style.
Quick question?? Hi thanks for your content. Im wondering if i buy a plan from Sudowrite do I have to purchase an Ai plan too? For example Claude and then plug it into Sudo? Or will Sudowrite have everything included? Ta.😮😮
In MidJourney, generate the character and refine it with any details you want included. Then use the character reference --cref to refer to this master image in later images. The format is --cref followed by the URL (internet address) where you have stored the image online. If generated in MidJourney, you can use the address where you created it in the first place, or you could use any hosting service. You can only use it for single characters currently, so if you have side characters in the image, you may want to add them in a separate step afterward, or they may all resemble the main character.
Thank you so much for this very excellent video. Thinking about writing my first children's novel beginning of the year 2025 and I was wondering how I was going to use midjourney and chat gpt to do it since I can't afford a real illustrator. This is excellent, exactly what I was looking for. Subscribed!!!
So, you want people to buy your book. (pay you for your work). but you wouldn't pay an artist for theirs? you can get a book cover for £50 easy. some will do work based on a profit share.. just need to actually look. instead of using a plagiarism machine an paying thieves (midjourney).
hey, are you planning to maybe make another video about novelai? they got a new model out right now, sadly its still only 8k but apparently its worth it if you know your way around the lorebooks and context optimization, which i dont, despite reading all the stuff in their reddit/discord threads ;-;
Hey there! I have followed your videos for a long time but I have a question; Have you done something with your sound lately? It seems to me that the loudness/sound compression is way more intense than before. This makes the sound/voice very hard for the ear to listen to unfortunately.
My story has disabled characters and I have trouble illustrating scenes. Villain has prosthetic arm. I cant get it work. I even tried justto ask it to make a gloved hand but no help.
We should talk, I built a gpt that gives reviews on chapters up to 5k words in the personalities of known popular A.I.'s such as Bender, HAL, Johnny 5, and so many more. It's awesome for scifi writers to play with. And kinda legit too. It wasn't nice to some of my work, I'll be honest
Jason, I'm not totally sure how best to get in touch with you outside of your membership, but I just wanted to let you know that I hope you're okay. It's not like you to go so long without posting a new video. I assume it's because you've got something big in the works, but just be aware people care about you and are wishing you well.
I was dealing with a medical issue. and was in the hospital for a bit. I’ve started producing videos again though it will be a few more days before they start posting.
Yeah, "young girl" seems to be code for prepubescent now. I see this in US news a lot too. Maybe it is like how the "young adult" audience has become younger and younger. "Teen girl" seems to work, but giving an exact age may be better.
I did an illustrated story (I'll add a link below) and getting the characters to look the same was a NIGHTMARE... but maybe it's because I wasn't using Midjourney and that was the mistake? th-cam.com/video/upq0zrfaXwc/w-d-xo.htmlsi=HgHaWigVKK7Zn3lf
Yeah, all of us are not wealthy. It takes work to make good AI art, more than you haters can imagine. Anyone can make A image, you have to have artistic eye to make a GOOD image.
@@RikuKiforForsstrom yeah, it takes work like typing four lines of text and hit "Retry" over and over again to reproduce art theft with glossy skin texture and Instagirl's facial expression😂
prompt:
you will be given a book chapter or article text your task is to select three to four key moments from the text or the main subject of an article and create visual descriptions of these scenes as image prompts for an AI image generator follow these steps :
1- read and analyze the following text you are given carefully take note of characters and location details.
2- step two select three to four key moments from the text these should be significant scenes or events that are crucial to the narrative or
content or they could be a singular character portrait.
3- for each key moment create a visual description to be used as an image prompt follow these guidelines for each prompt:
a- Begin with the most important part of the prompt (the main subject) near the beginning.
b- focus on one to two subjects maximum describing each separately.
c- include key details separated by commas The Prompt doesn't need to be
grammatically correct it can be a list of image details.
e- be specific and descriptive including details about setting character actions and atmosphere.
f- make sure the characters and locations include accurate descriptive details from the text and that these are consistent across your three to four prompts.
4- format your output as follows:
prompt 1:
[your first image prompt here]
prompt 2:
[your second image prompt here]
prompt 3:
[your third image prompt here]
5- end the prompt with the following " colouring book, simple illustration, fancy, historical novel style"
remember each prompt should be concise yet Vivid capturing the essence of the key moment in a way that can be visually represented, avoid abstract Concepts and focus on concrete visual elements.
Thank ypu bro
@@Truth-reins-free Left-click and hold the mouse until you highlight the entire text, then right-click on it and choose copy. I assume you are a typewriter person?
This is awesome stuff. One thing to get the red streak is to RUN her character first, to get her as you wanted , then use the text prompts with her as an Image Ref, and Say, "A bright red Colored Highlight streak in her hair that goes from the scalp of her hair to the tips" you can also add Style Ref images to the prompt if you want a consist painting style, that will maintain the same look throughout your book.
Wow!
This .
Newbie on this particular stuff. How would I get the system to look to an image that I have on my computer? Do I need to upload to a cloud, and set a URL for the link?
Great idea. I did just that a few months ago for a book cover to my dads memoirs when he was growing up in Poland. I took a whole chapter, reduced it with chatgpt, plug it in as a prompt and I get an awesome book cover the first attempt! Im also using my dads voice to make an audio version with elevenlabs. A very long ways from being done but its exciting.
Jason, you're killing it lately! Thanks for all you do
Great! I used a derivation in both Claude and Gemini and they worked really well. Gemini was able to send it right to image.
One thing I added was a step where I pick which setting to add to end to designate the style of image. Based on picking from a list, the AI added it efficiently and generated the image. An improvement I would love to add is a list of all the various settings that could be used for the style. Also, a selection prompt for things like orientation and any other choices that one would make to the image generator (such as Midjourney). Any idea where such a list may already exist?
Great stuff. I'd love to see a GPT o1 vs. Claude 3.5 comparison video as well!
I have been trying to figure out the best way to do this for days. This is perfect.
Great helpful video!
Thanks for this!
Does this prompt ensure character consistency over a range of images?
If you subscribe to a program like Flux or whatever, they have tools that allow you to do a face swap. Generate the image with the prompt, then swap the face with your character's.
@@Ank3rmanCan you do this for multiple characters?
I've been looking for a way to create several characters that I'd like to turn into a mixed reality educational cartoon, where the characters interact with my real world but I'm not sure we are quite there yet.
The characters would need to be consistent between episodes, and potentially overlap video or images in the real world.
@@tracy419 I believe so. Matt Wolfe has a recent video on AI he uses every day and he went through the process for creating a thumbnail in flux. You just paint over which face you want to swap and upload the reference image. It's definitely doable with today's technology but it might require just a little effort and know-how.
@@Ank3rman thanks, I'll check it out👍
Love it! Is there a copy of this prompt in its entirety somewhere to read over carefully? Thanks!
Here you go:
You will be given a book chapter or article text. Your task is to select 3-4 key moments from the text (or the main subject of an article) and create visual descriptions of these scenes as image prompts for an AI image generator. Follow these steps:
1. Read and analyze the following text you are given carefully. Take note of character and location details.
2. Select 3-4 key moments from the text. These should be significant scenes or events that are crucial to the narrative or content. Or they could be a singular character portrait.
3. For each key moment, create a visual description to be used as an image prompt. Follow these guidelines for each prompt:
a. Begin with the most important part of the prompt (the main subject) near the beginning.
b. Focus on 1-2 subjects maximum, describing each separately.
c. Include key details, separated by commas.
d. The prompt doesn't need to be grammatically correct; it can be a list of image details.
e. Be specific and descriptive, including details about setting, characters, actions, and atmosphere.
f. Describe characters without naming them (e.g., "a young man standing in a field," not "Tom standing in a field").
g. Include the general time period or setting, especially if it is not modern (e.g., "medieval village," "Victorian street").
h. Make sure the characters and locations include accurate descriptive details from the text and that these are consistent across your 3-4 prompts.
4. Format your output as follows:
Prompt 1:
[Your first image prompt here]
Prompt 2:
[Your second image prompt here]
Prompt 3:
[Your third image prompt here]
Prompt 4:
[Your fourth image prompt here]
5. End each prompt with the following:
"watercolor children's illustration, simple illustration, wonderous"
Remember, each prompt should be concise yet vivid, capturing the essence of the key moment in a way that can be visually represented. Avoid abstract concepts and focus on concrete, visual elements.
@@jackelmy392 Thank you! Much appreciated.
Nice tutorial, thank you so much, would be nice if you can share the fixed full prompt >>> 9:06
Here it is:
You will be given a book chapter or article text. Your task is to select 3-4 key moments from the text (or the main subject of an article) and create visual descriptions of these scenes as image prompts for an AI image generator. Follow these steps:
1. Read and analyze the following text you are given carefully. Take note of character and location details.
2. Select 3-4 key moments from the text. These should be significant scenes or events that are crucial to the narrative or content. Or they could be a singular character portrait.
3. For each key moment, create a visual description to be used as an image prompt. Follow these guidelines for each prompt:
a. Begin with the most important part of the prompt (the main subject) near the beginning.
b. Focus on 1-2 subjects maximum, describing each separately.
c. Include key details, separated by commas.
d. The prompt doesn't need to be grammatically correct; it can be a list of image details.
e. Be specific and descriptive, including details about setting, characters, actions, and atmosphere.
f. Describe characters without naming them (e.g., "a young man standing in a field," not "Tom standing in a field").
g. Include the general time period or setting, especially if it is not modern (e.g., "medieval village," "Victorian street").
h. Make sure the characters and locations include accurate descriptive details from the text and that these are consistent across your 3-4 prompts.
4. Format your output as follows:
Prompt 1:
[Your first image prompt here]
Prompt 2:
[Your second image prompt here]
Prompt 3:
[Your third image prompt here]
Prompt 4:
[Your fourth image prompt here]
5. End each prompt with the following:
"watercolor children's illustration, simple illustration, wonderous"
Remember, each prompt should be concise yet vivid, capturing the essence of the key moment in a way that can be visually represented. Avoid abstract concepts and focus on concrete, visual elements.
What if you run a local generator like ComfyUI for example? What model would be good to use to "try" and generate a novel style images from CivitAI? I have a ton of models so far that does SFW and NSFW of course. But, I'm willing to get a model that can do the style.
7:29 - The horse is actually melded into the wagon. lol.
Oooh, this seems like a great start to a book-to-graphic-novel adaptation.
Quick question?? Hi thanks for your content. Im wondering if i buy a plan from Sudowrite do I have to purchase an Ai plan too? For example Claude and then plug it into Sudo? Or will Sudowrite have everything included? Ta.😮😮
They have it included. But I prefer Novelcrafter.
One question: How do you keep the image of your character(s) consistent from chapter to chapter?
If you subscribe to a program like Flux or whatever, they have tools that allow you to do a face swap. Generate the image, then swap the face.
Midjourney lets you specify a reference image.
In MidJourney, generate the character and refine it with any details you want included. Then use the character reference --cref to refer to this master image in later images. The format is --cref followed by the URL (internet address) where you have stored the image online. If generated in MidJourney, you can use the address where you created it in the first place, or you could use any hosting service. You can only use it for single characters currently, so if you have side characters in the image, you may want to add them in a separate step afterward, or they may all resemble the main character.
@@MagnusItland Thanks - appreciate the feedback
@@LTHanlon Gracias
Thank you so much for this very excellent video. Thinking about writing my first children's novel beginning of the year 2025 and I was wondering how I was going to use midjourney and chat gpt to do it since I can't afford a real illustrator. This is excellent, exactly what I was looking for. Subscribed!!!
So, you want people to buy your book. (pay you for your work). but you wouldn't pay an artist for theirs? you can get a book cover for £50 easy. some will do work based on a profit share.. just need to actually look. instead of using a plagiarism machine an paying thieves (midjourney).
Please do a livestream with gpt 01!
This is gold! I have to search through 500 videos to find gems like this. Only ask is try to include us non fiction guys too 😁
Ie "this would apply to a non fiction book in this way..."
If you did an all encompassing "how to non fiction with ai" video, that would be a dream come true
hey, are you planning to maybe make another video about novelai? they got a new model out right now, sadly its still only 8k but apparently its worth it if you know your way around the lorebooks and context optimization, which i dont, despite reading all the stuff in their reddit/discord threads ;-;
Hey there!
I have followed your videos for a long time but I have a question;
Have you done something with your sound lately? It seems to me that the loudness/sound compression is way more intense than before. This makes the sound/voice very hard for the ear to listen to unfortunately.
My story has disabled characters and I have trouble illustrating scenes. Villain has prosthetic arm. I cant get it work. I even tried justto ask it to make a gloved hand but no help.
We should talk, I built a gpt that gives reviews on chapters up to 5k words in the personalities of known popular A.I.'s such as Bender, HAL, Johnny 5, and so many more. It's awesome for scifi writers to play with. And kinda legit too. It wasn't nice to some of my work, I'll be honest
Great to turn your book into a comic book.
Jason, I'm not totally sure how best to get in touch with you outside of your membership, but I just wanted to let you know that I hope you're okay. It's not like you to go so long without posting a new video. I assume it's because you've got something big in the works, but just be aware people care about you and are wishing you well.
I was dealing with a medical issue. and was in the hospital for a bit. I’ve started producing videos again though it will be a few more days before they start posting.
@@TheNerdyNovelist I'm glad you're out of the hospital!
Yeah, "young girl" seems to be code for prepubescent now. I see this in US news a lot too. Maybe it is like how the "young adult" audience has become younger and younger. "Teen girl" seems to work, but giving an exact age may be better.
I did an illustrated story (I'll add a link below) and getting the characters to look the same was a NIGHTMARE... but maybe it's because I wasn't using Midjourney and that was the mistake?
th-cam.com/video/upq0zrfaXwc/w-d-xo.htmlsi=HgHaWigVKK7Zn3lf
Unblock me coward why do you need to block my comment if you have nothing to fear?
Get a life, dude. There are millions of other videos you can watch.
Just say you're too broke to commission an artist
Yeah, all of us are not wealthy. It takes work to make good AI art, more than you haters can imagine. Anyone can make A image, you have to have artistic eye to make a GOOD image.
@@RikuKiforForsstrom yeah, it takes work like typing four lines of text and hit "Retry" over and over again to reproduce art theft with glossy skin texture and Instagirl's facial expression😂
@@ohiobumassit certainly isn’t theft how bout you do some research
@@edselsledge2940 loose the database that AI took without OG creators' permissions, let's see what it can do without stolen contents.
@@edselsledge2940 Sam Altman himself told press that without taking copyrighted content on the Internet his AI would not be possible.
Bro. Bro.. bruh. 🫡