Skip to main content 홈 크리에이터 jeremylongshore tons-of-skills-marketplace ideogram-reference-architecture
ideogram-reference-architecture Implement Ideogram reference architecture with prompt templates, asset pipelines, and CDN delivery.
Use when designing new Ideogram integrations, building brand asset systems,
or establishing architecture for image generation at scale.
Trigger with phrases like "ideogram architecture", "ideogram project structure",
"ideogram brand assets", "ideogram pipeline design", "ideogram at scale".
설치로 이동 Skills Marketplace 커뮤니티가 만든 AI 스킬을 발견하고 탐색하세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/jeremylongshore/tons-of-skills-marketplace --skill ideogram-reference-architecture명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
Zip 다운로드 다운로드 중... 이 저장소의 다른 Skills langchain-deploy-integration Deploy a LangChain 1.0 / LangGraph 1.0 app to Cloud Run, Vercel, or LangServe correctly — with timeouts sized for chain length, cold-start mitigation, SSE anti-buffering headers, and Secret Manager over .env. Use when prepping a first production deploy, debugging a stream that hangs behind a proxy, or diagnosing p99 latency spikes. Trigger with "langchain deploy", "langchain cloud run", "langchain vercel python", "langchain langserve", or "langchain docker".
langchain-langgraph-agents Build a correct LangGraph 1.0 ReAct agent with create_react_agent — typed tools, error propagation, recursion caps, and stop conditions that actually stop. Use when writing a first tool-calling agent, migrating from AgentExecutor or initialize_agent, or diagnosing an agent that loops on vague prompts. Trigger with "langgraph agent", "create_react_agent", "langgraph tool calling", "AgentExecutor migration", or "agent loop cost".
langchain-langgraph-human-in-loop Build LangGraph 1.0 human-in-the-loop approval flows with interrupt_before /
interrupt_after and Command(resume=...) — JSON-serializable state, clean
resume semantics, and UI wiring for approval decisions. Use when adding an
approval gate before an expensive tool call, wiring a Slack/web UI for agent
approvals, or debugging a graph that crashes on interrupt.
Trigger with "langgraph human in loop", "langgraph interrupt_before",
"langgraph approval flow", "Command resume", "langgraph HITL".
name ideogram-reference-architecture description Implement Ideogram reference architecture with prompt templates, asset pipelines, and CDN delivery.
Use when designing new Ideogram integrations, building brand asset systems,
or establishing architecture for image generation at scale.
Trigger with phrases like "ideogram architecture", "ideogram project structure",
"ideogram brand assets", "ideogram pipeline design", "ideogram at scale".
allowed-tools Read, Write, Edit, Grep version 1.10.0 license MIT author Jeremy Longshore <jeremy@intentsolutions.io> tags ["saas","ideogram","architecture","reference"] compatibility Designed for Claude Code
Ideogram Reference Architecture
Overview
Production architecture for AI image generation with Ideogram at scale. Covers prompt templating for brand consistency, generation pipelines using all six API endpoints, asset storage and CDN delivery, and metadata tracking for reproducibility.
Architecture Diagram
┌─────────────────────────────────────────────────────────┐
│ Prompt Engineering Layer │
│ Templates │ Brand Guidelines │ Negative Prompts │
└──────────────────────────┬──────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ Ideogram API (api.ideogram.ai) │
│ ┌──────────┐ ┌────────┐ ┌───────┐ ┌────────┐ │
│ │ Generate │ │ Edit │ │ Remix │ │Describe│ │
│ │(text→img)│ │(inpaint)│ │(vary) │ │(img→txt)│ │
│ └────┬─────┘ └───┬────┘ └──┬────┘ └───┬────┘ │
│ │ │ │ │ │
│ ┌────┴───────────┴─────────┴──────────┘ │
│ │ ┌──────────┐ ┌─────────┐ │
│ │ │ Upscale │ │ Reframe │ │
│ │ └────┬─────┘ └────┬────┘ │
│ └───────┴──────────────┘ │
└──────────────────────────┬──────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────────────┐
│ Post-Processing & Storage │
│ Download │ Resize │ WebP Convert │ S3/GCS │ CDN │
└─────────────────────────────────────────────────────────┘
Instructions
Step 1: Prompt Template System
interface PromptTemplate {
name : string ;
base : string ;
style : string ;
negativePrompt : string ;
aspectRatio : string ;
model : string ;
renderingSpeed ?: string ;
}
const BRAND_TEMPLATES : Record <string , PromptTemplate > = {
socialPost
name
"Social Media Post"
base
"{subject}, modern clean design, vibrant colors, professional"
style
"DESIGN"
negativePrompt
"blurry text, misspelled, watermark, low quality"
aspectRatio
"ASPECT_1_1"
model
"V_2"
blogHero
name
"Blog Hero Image"
base
"{subject}, editorial photography, wide composition, cinematic lighting"
style
"REALISTIC"
negativePrompt
"text overlay, watermark, blurry, oversaturated"
aspectRatio
"ASPECT_16_9"
model
"V_2"
storyVertical
name
"Story / Reel"
base
"{subject}, vertical composition, eye-catching, bold colors"
style
"DESIGN"
negativePrompt
"horizontal layout, small text, blurry"
aspectRatio
"ASPECT_9_16"
model
"V_2_TURBO"
ogImage
name
"Open Graph Image"
base
'{subject}, with text "{title}" in bold clean font, tech aesthetic'
style
"DESIGN"
negativePrompt
"blurry text, misspelled words, cluttered"
aspectRatio
"ASPECT_16_9"
model
"V_2"
function
buildPrompt
templateKey : string , vars : Record <string , string >
string
const
BRAND_TEMPLATES
if
throw
new
Error
`Unknown template: ${templateKey} `
let
base
for
const
of
Object
entries
replace
`{${key} }`
return
Step 2: Generation Service import { writeFileSync, mkdirSync } from "fs" ;
import { join } from "path" ;
const API_KEY = process.env .IDEOGRAM_API_KEY !;
async function generateFromTemplate (
templateKey : string ,
vars : Record <string , string >,
outputDir = "./assets"
) {
const template = BRAND_TEMPLATES [templateKey];
const prompt = buildPrompt (templateKey, vars);
const response = await fetch ("https://api.ideogram.ai/generate" , {
method : "POST" ,
headers : { "Api-Key" : API_KEY , "Content-Type" : "application/json" },
body : JSON .stringify ({
image_request : {
prompt,
model : template.model ,
style_type : template.style ,
aspect_ratio : template.aspectRatio ,
negative_prompt : template.negativePrompt ,
magic_prompt_option : "AUTO" ,
},
}),
});
if (!response.ok ) throw new Error (`Generate failed: ${response.status} ` );
const result = await response.json ();
const image = result.data [0 ];
const imgResp = await fetch (image.url );
const buffer = Buffer .from (await imgResp.arrayBuffer ());
mkdirSync (outputDir, { recursive : true });
const filename = `${templateKey} -${image.seed} .png` ;
writeFileSync (join (outputDir, filename), buffer);
return {
localPath : join (outputDir, filename),
seed : image.seed ,
prompt,
resolution : image.resolution ,
template : templateKey,
};
}
Step 3: Multi-Format Asset Pipeline import sharp from "sharp" ;
async function generateBrandAssetSet (subject : string , title : string const results = [];
for (const [key, template] of Object .entries (BRAND_TEMPLATES )) {
const asset = await generateFromTemplate (key, { subject, title });
results.push (asset);
await sharp (asset.localPath )
.webp ({ quality : 85 })
.toFile (asset.localPath .replace (".png" , ".webp" ));
await new Promise (r =>setTimeout (r, 3000 ));
}
const manifest = results.map (r =>template : r.template ,
seed : r.seed ,
prompt : r.prompt ,
files : {
png : r.localPath ,
webp : r.localPath .replace (".png" , ".webp" ),
},
}));
writeFileSync ("./assets/manifest.json" , JSON .stringify (manifest, null , 2 ));
console .log (`Generated ${results.length} brand assets with manifest` );
return results;
}
Step 4: Describe-then-Remix Pipeline
async function referenceBasedGeneration (referenceImagePath : string , modifications : string const form1 = new FormData ();
form1.append ("image_file" , new Blob ([readFileSync (referenceImagePath)]));
form1.append ("describe_model_version" , "V_3" );
const descResp = await fetch ("https://api.ideogram.ai/describe" , {
method : "POST" ,
headers : { "Api-Key" : API_KEY },
body : form1,
});
const descriptions = await descResp.json ();
const basePrompt = descriptions.descriptions [0 ].text ;
const form2 = new FormData ();
form2.append ("image" , new Blob ([readFileSync (referenceImagePath)]));
form2.append ("prompt" , `${basePrompt} , ${modifications} ` );
form2.append ("image_weight" , "40" );
form2.append ("rendering_speed" , "DEFAULT" );
const remixResp = await fetch ("https://api.ideogram.ai/v1/ideogram-v3/remix" , {
method : "POST" ,
headers : { "Api-Key" : API_KEY },
body : form2,
});
return remixResp.json ();
}
Project Structure project/
├── src/
│ ├── ideogram/
│ │ ├── client.ts # API wrapper
│ │ ├── templates.ts # Prompt templates
│ │ ├── pipeline.ts # Generation pipeline
│ │ └── types.ts # TypeScript types
│ ├── storage/
│ │ └── s3.ts # Image upload to S3/GCS
│ └── api/
│ └── generate.ts # API route handler
├── assets/ # Generated image output
│ └── manifest.json # Asset tracking
├── tests/
│ ├── templates.test.ts # Prompt template tests
│ └── pipeline.test.ts # Pipeline tests (mocked)
└── config/
├── ideogram.ts # API configuration
└── templates.json # Prompt templates (optional)
Error Handling Issue Cause Solution Inconsistent style No template system Use branded prompt templates URL expired Late download Download in same function call Text misspelled Prompt too vague Use DESIGN style, quote exact text Wrong aspect ratio Template mismatch Map templates to target platforms
Output
Prompt template system for brand consistency
Generation service with auto-download
Multi-format asset pipeline (PNG + WebP)
Describe-then-remix pipeline for reference-based generation
Asset manifest for tracking and reproducibility
Resources
Next Steps For multi-environment setup, see ideogram-multi-env-setup.
