Trigger Method
Triggered when the user says something like the following:
- "Make an image for this content / add a few images?"
- "Give me two (or more) image generation prompts"
- "Too much text looks bad, help me make it more fun and easier to read"
- "Package this process into a prompt template/skills"
- "/image " "/illustrate" "/generate image"
Process Overview
| Stage | Name | Goal | Detailed File |
|---|---|---|---|
| 1 | Requirements Clarification (Spec/DoD) | First dig into requirements: content/scenario/audience/text volume; default size and style; produce a one-sentence restatement and requirements summary | stages/01-brief.md |
| 2 | Image Planning (Break into Blocks -> Checklist) | Break down content -> define image checklist (how many images/what each covers/which template to use) | stages/02-plan.md |
| 3 | Copy Finalization (Copy Spec) | Finalize word by word "what is written on the image" (single source of truth) | stages/03-copy.md |
| 4 | Prompt Packaging (Prompt Pack) | Package the Copy Spec into copyable prompts; generate a batch request package and batch-generate images after user confirmation | stages/04-prompts.md |
| 5 | Iterative Refinement | Based on feedback, reduce text, change metaphors, improve readability | stages/05-iterate.md |
Scheduling Rules
How to determine the current stage:
- Requirements are not yet clearly stated (content + scenario + audience + more text/less text) -> Stage 1
- The article is very long and needs to be broken into blocks, or it is necessary to determine "how many images/what each covers" -> Stage 2
- The image checklist is confirmed, but "what is written word by word on the image" is not yet determined -> Stage 3
- The Copy Spec is confirmed and copyable prompts need to be produced; (optional) batch image generation after prompt confirmation -> Stage 4
- User feedback says "too much text/looks bad/does not fit the cover" -> Stage 5 (if necessary, roll back to Stage 1 to re-lock requirements and more text/less text)
At the start of each stage:
- Tell the user the current stage and this stage's output
- Read the corresponding stage file and execute step by step
Output Specifications (Must Be Followed)
- One "core message" per image; do not stuff explanatory paragraphs into the image
- All Chinese must be clearly readable: large font size, few words and short sentences, avoid dense small text
- Output each prompt in a separate code block for easy copying
- Default output is 16:9 landscape (unless the user explicitly wants 3:4 comic/vertical format)
- Default style: cream paper background + colored pencil watercolor hand-drawn + light doodles, fun but clean (can use
templates/style-block.md） - Once the Stage 3 output (Copy Spec) is confirmed, Stage 4 must not alter the copy without authorization; only package it and fill in parameters/constraints
Quick Start (Minimum Input for the User)
The user only needs to provide these four items to begin:
- The content to be illustrated (can be a paragraph, a subsection, or an entire article)
- Where it will be used + viewing distance (PPT projection viewed from afar / mobile phone viewed up close / poster)
- Who will view it (beginner/practitioner/boss/student...)
- Preference: more "clean with fewer words" or more "information density"
Optional additions (fine if you don't include them):
- What kind of images you probably want: cover/table of contents, single-page overview, lecture explanation, social media poster (if unsure, I'll recommend based on the scenario and preferences)
The delivery order you should follow:
- First output: image list (how many + one-sentence purpose for each + template suggestions) (Stage 2)
- After user confirmation: output Copy Spec one by one (word-for-word final draft) (Stage 3)
- After Copy Spec confirmation: output copyable prompts/call packages one by one (Stage 4)
- (Optional) After prompt confirmation: batch call the API to generate images (continuous process within Stage 4)
- If the user says "too much text/doesn't look good," enter iteration (Stage 5)
File structure
stages/
├── 01-brief.md
├── 02-plan.md
├── 03-copy.md
├── 04-prompts.md
└── 05-iterate.md
templates/
├── style-block.md
├── 16x9-infographic.md
├── 16x9-contrast-2cards.md
├── 16x9-3cards-insights.md
├── 16x9-cover-roadmap.md
├── 16x9-5panel-comic.md
├── api-config.md
├── apimart-curl.md
├── apimart-requests-jsonl.md
└── checklist.md
examples/
└── ai-tools-selection.md
scripts/
├── apimart_batch_generate.py
├── apimart.env.example
└── README.md