Reading note
This chapter collects creators' published production cases, emphasizing Jimeng. Each entry identifies the creator, tutorial, actual tools and where to obtain assets or prompts. Only sequences explicitly documented in the sources are summarized; follow the original tutorials for complete operation.
| Route | Original creator and existing case | What you can find directly |
|---|---|---|
| Jimeng Agent: start here | AI Product Sniper / Kevin Zhu: Mid-Autumn animation | Illustrated tutorial, Jimeng work, shot descriptions and a “Make similar” entry point |
| Jimeng Smart Multiframe | AI Product Sniper / Kevin Zhu: drink product showcase | Creator's demonstration and using multiple images as keyframes |
| Jimeng Seedance 2.0 | HiShineAi: grid-storyboard workflow for a racing short | Original video, tool combination and storyboard-to-video process |
| Wan 3.0 | The Workflow Lab / Terence Sia: replace your video character | Illustrated steps, screenshots, editable prompts and final output |
| MiniMax H3 + Seedance 2.5 | Curious Refuge / Tyler Smith: drafts before final output | Actual iterations, references, before/after outputs and a full illustrated tutorial |
Keep the authors' versions and interfaces: the Jimeng cases include older models; the Wan and H3 cases use Higgsfield and Magnific, respectively. They are not button-by-button tutorials for the Wan or Hailuo original sites, and the newer model names in Chapter 1 cannot simply be substituted. A readable tutorial does not mean all assets are public; see each entry's access conditions.
1. Jimeng Agent: reuse the creator's Mid-Autumn animation
Original creator: AI Product Sniper / Kevin Zhu. Published October 9, 2025.
Tutorial and work: “Generate a Mid-Autumn animation in one minute! Jimeng Agent in practice!”; creator's Jimeng work and Make similar entry.
The creator used Jimeng Agent for a family-reunion animation and supplied a link to the work. The public storyboard runs from returning home with a lantern through a family reunion to moon viewing in a courtyard. The page labels it Seedance 1.0 mini, 42 seconds, 16:9.
Creator's process: describe the animation to the agent, generate a multishot work, then use the shared work to begin a similar creation.
How to reuse this existing process: read the tutorial, inspect each shot on the work page, then use “Make similar” to enter the platform's reuse flow and follow the original instructions. Retain the storyboard structure before adjusting theme and content.
Assets and version conditions: the original page exposes shot images and video prompts. The complete initial request to the agent was not obtained in this research, so no substitute is written here. Confirm the editable scope and continued availability of the older model after logging in.
2. Jimeng Smart Multiframe: reuse the drink showcase method
Original creator: AI Product Sniper / Kevin Zhu. Published August 20, 2025.
Original tutorial: “Create an explosive product showcase in one click! Jimeng practice, episode 27!”.
The creator demonstrated a fictional drink of their own design with Jimeng's then-current Smart Multiframe feature.
Creator's process: prepare several product images as keyframes and describe the transition from each frame to the next to assemble the showcase.
How to reuse this existing process: study the image ordering and transition descriptions, then substitute your product images. Reuse the keyframe organization method; identify models and controls from the original tutorial rather than calling this a Seedance 2.5-specific workflow.
Asset conditions: the inspected article did not publish an original product-image package, complete interframe prompts or fixed generation settings. This is a demonstration tutorial, not a downloadable one-click workflow.
3. Jimeng Seedance 2.0: follow the racing storyboard workflow
Original creator: HiShineAi. Published February 27, 2026.
Original video tutorial: “10× efficiency! A workflow for AI shorts using nine-panel storyboards”.
The creator demonstrates the overall process with a racing short, listing Gemini, Nano Banana Pro and Jimeng Seedance 2.0 as the core tools.
Creator's process: organize shot logic in Gemini, establish consistent visuals through nine- or twenty-five-panel grids, then combine Seedance 2.0 and individual-frame shots into a short.
How to reuse this existing process: follow the original video to learn how storyboard planning, grid images and individual shots work together. Keep the creator's tool combination for a first attempt, then change the subject; do not replace the storyboard method with an invented single prompt.
Asset conditions: the creator's description says prompt templates require contacting them as instructed; there is no public direct download. This research verified the original account, case description and workflow summary, but did not obtain full subtitles, step-by-step parameters or template text. Consult the video for those details; they are not filled in here.
4. Wan 3.0: retain your performance while changing the character
Original creator: The Workflow Lab / Terence Sia. The tutorial's stated verification date is October 23, 2026.
Original tutorial: “Transform yourself with Wan 3.0”; about the creator; Instagram demonstration linked by the tutorial.
This tutorial actually uses Wan 3.0 inside Higgsfield and discloses a partnership with Higgsfield. It includes interface screenshots, a character reference, prompts and the final output.
Creator's process:
- Record your performance, retaining movement, expressions and sound.
- Create a character image using the provided character-reference prompt.
- Select Wan 3.0 in Higgsfield under Video → Create Video.
- Upload the character image and performance video under References and check their identifiers.
- Copy Your editable transformation prompt from the original article and replace fields such as duration and aspect ratio.
- Set the output, enable sound, generate and inspect character, motion and audio, then download.
Obtaining the prompts: use the two editable templates in the original tutorial. It also lists the creator's actual longer prompt, including the extra action of adding a coffee cup at a specified time. The author explicitly distinguishes it from the simpler reusable template.
5. MiniMax H3: approve a draft before generating the final
Original creator: Curious Refuge / Tyler Smith. The original page displays October 9.
Original tutorial: “How to Save Money on AI Video Generation”.
The creator made a three-shot bar conversation in Magnific, publishing character images, scene references, multiple iterations and the final assembly. This section covers the MiniMax H3 draft → Seedance 2.5 final route.
Creator's process:
- Divide the three shots into two parts: combine the first two into approximately ten seconds and make the third separately.
- Upload character and scene references; iterate on composition and staging with low-resolution MiniMax H3 Max outputs.
- Once the direction is approved, use the revised prompt in Seedance 2.5 standard mode for the final version.
- For the third shot, add the previously generated clip as a continuity reference, then assemble the results.
Assets and usage conditions: the article includes references and iteration results but does not publish the full generation prompt for this case; no template is invented here. Its cost, time and quality judgments come from that Magnific test, not a Hailuo-site price or quality promise.
The article also tests Seedance's own Draft → HD route and compares the methods. Read that comparison if you want fewer platform and model switches.
6. Choose how to reuse an original tutorial
| How you want to start | Corresponding original tutorial |
|---|---|
| Open an existing work and make a similar one | Mid-Autumn Agent animation: includes a Jimeng work entry point |
| Make a showcase with your product images | Drink Smart Multiframe: study keyframes and transitions |
| Learn to organize a multishot short | HiShineAi racing short: follow the original video |
| Obtain publicly available editable prompts | Terence Sia's Wan 3.0 character-transformation tutorial |
| Study generation iterations and final-output costs | Tyler Smith's H3 drafts to Seedance final tutorial |
This chapter checked creator pages, original articles or public descriptions and links, but has not reproduced the cases in accounts. Original assets, prompts, parameters or subtitles that were not published are not reverse-engineered from finished videos and invented here.