above the embers
above the embers
- By
- Wei
- Steps
- 58
- Connections
- 74
- Updated
- 15 Sept 2026
Inside this workflow
30 of 43 runs
Each run is the workflow’s own record: the model it called, the prompt it was given, and what came out.
- 01
Image generator
Image generatorNano Banana Pro- Input
Character height ratio-3D
Su Wan-1 (3D)
Linye-1 (3D)
Lao Pao-1 (3D)
PromptCompare four characters {{Image 1}} {{Image 2}} {{Image 3}} {{Image 4}} with reference to the height ratio {{Image 5}} to generate a set of character combinations with completely different clothing, weapons, action poses, etc. Add the combined name "ABOVE THE ASHES" behind the character, in English font, bold and powerful, in large characters. Avoid messy details in the generation, and the picture should be soft to make the characters appear more clean and three-dimensional.
- 02
Video generator
Video generatorSeedance 2.0- Input
01Go to run 01
Forestry-3D three views
PromptDark cinematic masterpiece, hyper-realistic CG rendering, Unreal Engine 5 style, extreme bullet time, ultra-slow motion, mysterious and ethereal atmosphere, elusive mood, high contrast lighting (chiaroscuro), dense volumetric fog, dark steampunk/dieselpunk aesthetic, 8k resolution, extreme micro-details (visible skin pores, fine metal scratches, rusted iron, worn leather textures), subsurface scattering on skin, dramatic shadows, cinematic color grading. ### **The first paragraph (0s - 15s): Roaring embers (focus: heavy gunners and wasteland atmosphere)**
Output1 output — open the canvas to view
- 03
Video generator
Video generatorSeedance 2.0- Input
PromptDark cinematic masterpiece, hyper-realistic CG rendering, Unreal Engine 5 style, extreme bullet time, ultra-slow motion, mysterious and ethereal atmosphere, elusive mood, high contrast lighting (chiaroscuro), dense volumetric fog, dark steampunk/dieselpunk aesthetic, 8k resolution, extreme micro-details (visible skin pores, fine metal scratches, rusted iron, worn leather textures), subsurface scattering on skin, dramatic shadows, cinematic color grading.****The Ritual of the Recorder in White and the Radiated Echo****\ Fading out from the darkness of the gun chamber, a beam of cold white light fell on the one from the left
Output1 output — open the canvas to view
- 04
Video generator
Video generatorSeedance 2.0- Input
PromptDark cinematic masterpiece, hyper-realistic CG rendering, Unreal Engine 5 style, extreme bullet time, ultra-slow motion, mysterious and ethereal atmosphere, elusive mood, high contrast lighting (chiaroscuro), dense volumetric fog, dark steampunk/dieselpunk aesthetic, 8k resolution, extreme micro-details (visible skin pores, fine metal scratches, rusted iron, worn leather textures), subsurface scattering on skin, dramatic shadows, cinematic color grading.******Wild Girl and Broken Time******\ White light cuts into the interior of a dark abandoned building, the camera is at a low angle
Output1 output — open the canvas to view
- 05
Video generator
Video generatorSeedance 2.0- Input
PromptDark cinematic masterpiece, hyper-realistic CG rendering, Unreal Engine 5 style, extreme bullet time, ultra-slow motion, mysterious and ethereal atmosphere, elusive mood, high contrast lighting (chiaroscuro), dense volumetric fog, dark steampunk/dieselpunk aesthetic, 8k resolution, extreme micro-details (visible skin pores, fine metal scratches, rusted iron, worn leather textures), subsurface scattering on skin, dramatic shadows, cinematic color grading.********Knife shadow, gunfire and the rising title of the film********\ After the debris scene transitioned, a tall man on the right stood
Output1 output — open the canvas to view
- 06
Video generator
Video generatorSeedance 2.0- Input
Su Wan-3D three views
PromptDark cinematic masterpiece, hyper-realistic CG rendering, Unreal Engine 5 style, extreme bullet time, ultra-slow motion, mysterious and ethereal atmosphere, elusive mood, high contrast lighting (chiaroscuro), dense volumetric fog, dark steampunk/dieselpunk aesthetic, 8k resolution, extreme micro-details (visible skin pores, fine metal scratches, rusted iron, worn leather textures), subsurface scattering on skin, dramatic shadows, cinematic color grading. ### **Paragraph 2 (1s - 5s): The Weight of Truth (Focus: Scholars and Mysterious Books)** -
Output1 output — open the canvas to view
- 07
Video generator
Video generatorSeedance 2.0- Input
PromptDark cinematic masterpiece, hyper-realistic CG rendering, Unreal Engine 5 style, extreme bullet time, ultra-slow motion, mysterious and ethereal atmosphere, elusive mood, high contrast lighting (chiaroscuro), dense volumetric fog, dark steampunk/dieselpunk aesthetic, 8k resolution, extreme micro-details (visible skin pores, fine metal scratches, rusted iron, worn leather textures), subsurface scattering on skin, dramatic shadows, cinematic color grading. ### **The third paragraph (1s - 15s): The cold light that breaks the fog (focus: assassins and swords)** - *
Output1 output — open the canvas to view
- 08
Video generator
Video generatorSeedance 2.0- Input
PromptDark cinematic masterpiece, hyper-realistic CG rendering, Unreal Engine 5 style, extreme bullet time, ultra-slow motion, mysterious and ethereal atmosphere, elusive mood, high contrast lighting (chiaroscuro), dense volumetric fog, dark steampunk/dieselpunk aesthetic, 8k resolution, extreme micro-details (visible skin pores, fine metal scratches, rusted iron, worn leather textures), subsurface scattering on skin, dramatic shadows, cinematic color grading. ### **The fourth paragraph (46s - 60s): Oath above the ashes (emphasis: heavy swordsmen and group portraits)*
Output1 output — open the canvas to view
- 09
Video generator
Video generatorSeedance 2.0- Input
PromptDark cinematic masterpiece, hyper-realistic CG rendering, Unreal Engine 5 style, extreme bullet time, ultra-slow motion, mysterious and ethereal atmosphere, elusive mood, high contrast lighting (chiaroscuro), dense volumetric fog, dark steampunk/dieselpunk aesthetic, 8k resolution, extreme micro-details (visible skin pores, fine metal scratches, rusted iron, worn leather textures), subsurface scattering on skin, dramatic shadows, cinematic color grading.**Appearance from the Ashes**\ In the black field, a grain of ash spins and falls at an extremely slow speed. The camera moves forward in extreme close-up, and the grain of ash appears.
Output1 output — open the canvas to view
- 10
Video generator
Video generatorSeedance 2.0- Input
PromptDark cinematic masterpiece, hyper-realistic CG rendering, Unreal Engine 5 style, extreme bullet time, ultra-slow motion, mysterious and ethereal atmosphere, elusive mood, high contrast lighting (chiaroscuro), dense volumetric fog, dark steampunk/dieselpunk aesthetic, 8k resolution, extreme micro-details (visible skin pores, fine metal scratches, rusted iron, worn leather textures), subsurface scattering on skin, dramatic shadows, cinematic color grading.********Knife shadow, gunfire and title freeze frame********\ After the debris scene transitioned, a tall man on the right stood
Output1 output — open the canvas to view
- 11
Video generator
Video generatorSeedance 2.0- Input
PromptDark cinematic masterpiece, hyper-realistic CG rendering, Unreal Engine 5 style, extreme bullet time, ultra-slow motion, mysterious and ethereal atmosphere, elusive mood, high contrast lighting (chiaroscuro), dense volumetric fog, dark steampunk/dieselpunk aesthetic, 8k resolution, extreme micro-details (visible skin pores, fine metal scratches, rusted iron, worn leather textures), subsurface scattering on skin, dramatic shadows, cinematic color grading.****The Ceremony of White Clothes and Radiant Echo****\ Fading out from the darkness of the gun chamber, a beam of cold white light fell on the second white man from the left.
Output1 output — open the canvas to view
- 12
Image generator
Image generatorNano Banana Pro- Input
Xiaoan-1 (3D)
PromptCompare four characters {{Image 1}} {{Image 2}} {{Image 3}} {{Image 4}} with reference to the height ratio {{Image 5}} to generate a set of character combinations with completely different clothing, weapons, action poses, etc. Add the combined name "ABOVE THE ASHES" behind the character, in English font, bold and powerful, in large characters. Avoid messy details in the generation, and the picture should be soft to make the characters appear more clean and three-dimensional.
- 13
Image generator
Image generatorNano Banana ProPromptAccording to the height ratio of four characters, a set of character combination diagrams with completely different weapons, action postures, etc. are generated. Add the combined name "ABOVE THE ASHES" behind the character, in English font, bold and powerful, in large characters. Avoid messy details in the generation, and the picture should be soft to make the characters appear more clean and three-dimensional. The character's movements are exaggerated and tense but conform to physical logic and animation principles. The limbs are natural and dexterous, without deformation or collapse.
- 14
Image generator
Image generatorNano Banana ProPromptAccording to the height ratio of four characters, a set of character combination diagrams with completely different weapons, action postures, etc. are generated. Add the combined name "ABOVE THE ASHES" behind the character, in English font, bold and powerful, in large characters. Avoid messy details in the generation, and the picture should be soft to make the characters appear more clean and three-dimensional.
- 15
Image generator
Image generatorNano Banana Pro- Input
PromptThe four characters {{Image 1}} {{Image 2}} {{Image 3}} {{Image 4}} are generated according to the height ratio {{Image 5}} to generate a set of character combinations with completely different weapons, action postures, etc. {{Image 1}} is good at using rifles, {{Image 2}} is good at medicine configuration, {{Image 3}} is good at detection and traps, {{Image 4}} is proficient with knives and heavy shotguns. Add the combined name "ABOVE THE ASHES" behind the character, in English font, bold and powerful, in large characters. Avoid messy details in the generation, and the picture should be soft to make the characters appear more clean and three-dimensional. The character's movements are exaggerated and tense but conform to physical logic and animation principles. The limbs are natural and dexterous, without deformation or collapse.
- 16
Image generator
Image generatorNano Banana Pro- Input
Character height ratio
PromptCombine four characters {{Image 1}} {{Image 2}} {{Image 3}} {{Image 4}} according to the height ratio {{Image 5}} to generate a set of character combination diagrams with completely different character clothing, weapons, action poses, etc. Add the combined name "ABOVE THE ASHES" behind the character, in English font, bold and powerful, in large characters. Avoid messy details in the generation, and the picture should be soft to make the characters appear more clean and three-dimensional.
- 17
Image generator
Image generatorNano Banana Pro- Input
- 18
Image generator
Image generatorNano Banana ProPromptAccording to the height ratio of four characters, a set of character combination diagrams with completely different weapons, action postures, etc. are generated. Add the combined name "ABOVE THE ASHES" behind the character, in English font, bold and powerful, in large characters. Avoid messy details in the generation, and the picture should be soft to make the characters appear more clean and three-dimensional. The character's movements are exaggerated and tense but conform to physical logic and animation principles. The limbs are natural and dexterous, without deformation or collapse.
- 19
Text generator
Text generatorChat gpt-5.5- Input
PromptBased on this picture, refer to the different characters and their weapon attributes in the picture, generate a 15-second video prompt where the characters appear one by one and finally gather together, requiring each character to have a handsome personal display and corresponding weapon usage skills (at least three storyboard performances for each character, one close-up, one medium shot, and one panorama), and the appearance of the characters The connection of modes must be smooth and cool, seamless transitions (such as masking transitions, dynamic transitions, similar transitions, seamless transitions or corresponding weapon usage skills as the connection method to avoid too many irrelevant hard-cut shots). Special lens expression techniques must be used, and the connection and switching of pictures can be achieved through multi-storyboards, multi-views, and multi-special angles. The overall rhythm is compact and orderly, the logic is clear, the characters' movements are exaggerated and tense, and the coordination with the camera is extremely visual tension. It is of the quality of a top-notch CG animated film. At the end of the film, each character appears handsomely in the position of the reference picture one by one and keeps moving (accompanied by the rapid appearance of background text one stroke at a time). The movements of the characters are correct and consistent with physical logic and animation principles. The movements are deft, vivid and exaggerated without distortion, collapse, confusion and other problems. The entire film is free to move, including circling, rotating, pushing, pulling, panning and other difficult combinations of camera movements (the whole film has a tight rhythm and a dynamic feel that is overwhelming). The whole thing needs to be combined with some changes in fast and slow shots. The scenes vary widely, ranging from extreme close-ups to panoramic displays. There is no background music or subtitles in the entire film.
Output1 output — open the canvas to view
- 20
Text generator
Text generatorChat gpt-5.5- Input
PromptI want to use this reference image to generate a 15-second video, one shot per second, in slow motion (can be used as a storyboard). Because it is used as a storyboard, the overall style must remain the same, and the composition of each picture must be extremely professional. The tone, light and shadow, element combination, and character scenes all reach the top CG hyper-realistic style. Every picture is flawless. The characters are extremely detailed and real. There need to be different sceneries, different compositions, different characters, and scenes corresponding to the same world view. Characters can appear alone, together, or as separate empty shots of scenes. Focus on expressing each person's characteristics and their unique weapon skills. Each picture needs to have a certain correlation, and will eventually be used as the opening title of a film or television drama. Write a complete prompt word, which can be perfectly generated directly with seedance2.0.
Output1 output — open the canvas to view
- 21
Text generator
Text generatorChat gpt-5.5- Input
PromptBased on the characters in the reference picture and the attributes of the weapons they use, an opening sequence is generated for the opening of the film. The duration of the film is 60 seconds, and the appearance time of each character is relatively balanced. All are in slow motion (bullet time, Bullet Time), including the performance of character details, the use of weapons in slow motion, empty shots of the scene atmosphere, and shots showing the weapons alone. The entire film is full of slow camera movements (including zoom shots, pull shots, surround shots, panning shots, etc.). Some shots require transitions or transitions between shots, and some are directly cut. The picture style is mysterious, ethereal, elusive, high contrast, foggy, film and television-level picture texture, extremely realistic and delicate CG realistic performance. Whether it’s the character’s skin or the detailed textures of weapons and scenes. There are many different scenes (including extreme close-up, close-up, panorama, medium shot, long shot), and various composition angles (center composition, Dutch angle composition, triangle cutout, leading line composition, etc.). The entire film needs to be story-telling, but the logic does not need to be coherent, and it can also be expressed in multiple threads. Write 4 short paragraphs, each paragraph contains 15 seconds of content
Output1 output — open the canvas to view
- 22
Text generator
Text generatorChat gpt-5.5- Input
PromptBased on this picture, refer to the different characters and their weapon attributes in the picture, generate a 15-second video prompt where the characters appear one by one and finally gather together, requiring each character to have a handsome personal display and corresponding weapon usage skills (at least three storyboard performances for each character, one close-up, one medium shot, and one panorama), and the appearance of the characters The connection of modes must be smooth and cool, seamless transitions (such as masking transitions, dynamic transitions, similar transitions, seamless transitions or corresponding weapon usage skills as the connection method to avoid too many irrelevant hard-cut shots). Special lens expression techniques must be used, and the connection and switching of pictures can be achieved through multi-storyboards, multi-views, and multi-special angles. The overall rhythm is compact and orderly, the logic is clear, the characters' movements are exaggerated and tense, and the coordination with the camera is extremely visual tension. It is of the quality of a top-notch CG animated film. At the end of the film, each character appears handsomely in the position of the reference picture one by one and keeps moving (accompanied by the rapid appearance of background text one stroke at a time). The movements of the characters are correct and consistent with physical logic and animation principles. The movements are deft, vivid and exaggerated without distortion, collapse, confusion and other problems. The entire film is free to move, including circling, rotating, pushing, pulling, panning and other difficult combinations of camera movements (the whole film has a tight rhythm and a dynamic feel that is overwhelming). The whole thing needs to be combined with some changes in fast and slow shots. The scenes vary widely, ranging from extreme close-ups to panoramic displays. There is no background music or subtitles in the entire film.
Output1 output — open the canvas to view
- 23
Text generator
Text generatorChat gpt-5.5- Input
PromptBased on this picture, refer to the different characters and their weapon attributes in the picture, generate a 15-second video prompt where the characters appear one by one and finally gather together, requiring each character to have a handsome personal display and corresponding weapon usage skills (at least three storyboard performances for each character, one close-up, one medium shot, and one panorama), and the appearance of the characters The connection of modes must be smooth and cool, seamless transitions (such as masking transitions, dynamic transitions, similar transitions, seamless transitions or corresponding weapon usage skills as the connection method to avoid too many irrelevant hard-cut shots). Special lens expression techniques must be used, and the connection and switching of pictures can be achieved through multi-storyboards, multi-views, and multi-special angles. The overall rhythm is compact and orderly, the logic is clear, the characters' movements are exaggerated and tense, and the coordination with the camera is extremely visual tension. It is of the quality of a top-notch CG animated film. At the end of the film, each character appears handsomely in the position of the reference picture one by one and keeps moving (accompanied by the rapid appearance of background text one stroke at a time). The movements of the characters are correct and consistent with physical logic and animation principles. The movements are deft, vivid and exaggerated without distortion, collapse, confusion and other problems. The entire film is free to move, including circling, rotating, pushing, pulling, panning and other difficult combinations of camera movements (the whole film has a tight rhythm and a dynamic feel that is overwhelming). The whole thing needs to be combined with some changes in fast and slow shots. The scenes vary widely, ranging from extreme close-ups to panoramic displays. There is no background music or subtitles in the entire film.
Output1 output — open the canvas to view
- 24
Text generator
Text generatorChat gpt-5.5PromptAdjust the prompt word of the second woman in the white dress to match her exclusive props: an old notebook and a radiation detector on her body, or the suggested props related to the potion bottle. {{Text 1}}
Output1 output — open the canvas to view
- 25
Text generator
Text generatorChat gpt-5.5- Input
PromptBased on the characters in the reference picture and the attributes of the weapons they use, an opening sequence is generated for the opening of the film. The film is 60 seconds long, all in slow motion (bullet time, Bullet Time), including character details, use of weapons in slow motion, empty shots of scene atmosphere, and shots showing weapons alone. The entire film is full of slow camera movements (including zoom shots, pull shots, surround shots, panning shots, etc.). Some shots require transitions or transitions between shots, and some are directly cut. The picture style is mysterious, ethereal, elusive, high contrast, foggy, film and television-level picture texture, extremely realistic and delicate CG realistic performance. Whether it’s the character’s skin or the detailed textures of weapons and scenes. There are many different scenes (including extreme close-up, close-up, panorama, medium shot, long shot), and various composition angles (center composition, Dutch angle composition, triangle cutout, leading line composition, etc.). The entire film needs to be story-telling, but the logic does not need to be coherent, and it can also be expressed in multiple threads. Write 4 short paragraphs, each paragraph contains 15 seconds of content
Output1 output — open the canvas to view
- 26
Text generator
Text generatorChat gpt-5.5- Input
PromptBased on the characters in the reference picture and the attributes of the weapons they use, an opening sequence is generated for the opening of the film. The film is 60 seconds long, all in slow motion (bullet time, Bullet Time), including character details, use of weapons in slow motion, empty shots of scene atmosphere, and shots showing weapons alone. The entire film is full of slow camera movements (including zoom shots, pull shots, surround shots, panning shots, etc.). Some shots require transitions or transitions between shots, and some are directly cut. The picture style is mysterious, ethereal, elusive, high contrast, foggy, film and television-level picture texture, extremely realistic and delicate CG realistic performance. Whether it’s the character’s skin or the detailed textures of weapons and scenes. There are many different scenes (including extreme close-up, close-up, panorama, medium shot, long shot), and various composition angles (center composition, Dutch angle composition, triangle cutout, leading line composition, etc.). The entire film needs to be story-telling, but the logic does not need to be coherent, and it can also be expressed in multiple threads. Write 4 short paragraphs, each paragraph contains 15 seconds of content
Output1 output — open the canvas to view
- 27
Text generator
Text generatorChat gpt-5.5- Input
[Scenario] World view-full series
PromptBased on this picture, with reference to the landforms and architectural features of different places in the picture, a 15-second video prompt is generated that summarizes the wasteland environment. Each scene is required to show the corresponding architectural details (each scene shows a panorama and a close-up), and the connection between scenes Seamless transitions must be achieved through transition methods (such as masking transitions, dynamic transitions, similar transitions, seamless transitions, and avoiding too many irrelevant hard cuts). Special lens expression techniques must be used, and the transition and switching of images can be achieved through multiple storyboards, multiple scenes, and multiple special angles. The overall rhythm is tight and orderly, the logic is clear, and the quality of a top-notch CG animated movie. The film finally focuses on the mysterious building complex in the distance. There are firelights mixed with twisted figures in the smoke. The thick smoke and firelight blur the outline of the Cthulhu-style monster parts, suggesting that endless crises are waiting. The actions of the characters are correct and consistent with physical logic and animation principles. The actions are deft, vivid and exaggerated without distortion, collapse, confusion and other problems. The entire film is free to move, including circling, rotating, pushing, pulling, panning and other difficult combinations of camera movements (the whole film has a tight rhythm and a dynamic feel that is overwhelming). The whole thing needs to be combined with some changes in fast and slow shots. The scenes vary widely, ranging from extreme close-ups to panoramic displays. There is no background music or subtitles in the entire film.
Output1 output — open the canvas to view
- 28
Text generator
Text generatorChat gpt-5.5PromptAdjust the last paragraph of prompt words so that the final frame can perfectly restore the content of Figure 1. {{Text 1}}
Output1 output — open the canvas to view
- 29
Text generator
Text generatorChat gpt-5.5PromptReferring to the characteristics of the second character from the left, adjust the second paragraph of text description to make it fit the props held by the character (the character holds an old notebook recording experimental data in his arms, and hangs an old-fashioned radiation detector on his body) {{Text 1}}
Output1 output — open the canvas to view
- 30
Video generator
Video generatorSeedance 2.0Prompt{{Text 1}}
Output1 output — open the canvas to view
The remaining 13 runs are on the canvas in CinAI Studio.