agentsclimarketplace

Ideogram sdk patterns

Skill jeremylongshore/claude-code-plugins-plus-skills/plugins/saas-packs/ideogram-pack/skills/ideogram-sdk-patterns

'Apply production-ready Ideogram API patterns for TypeScript and Python.From its SKILL.md

Install
npx -y skills add jeremylongshore/claude-code-plugins-plus-skills --skill ideogram-sdk-patterns

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its file declares

Copied from the file, not written here

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

8.0 KB, ~1.8k tokens by cl100k_base, as published. Nobody here has run it

Ideogram SDK Patterns

Overview

Production-ready patterns for Ideogram's REST API. Since Ideogram has no official SDK, these patterns provide type-safe wrappers, retry logic, response validation, and multi-tenant support for the api.ideogram.ai endpoints.

Prerequisites

  • Completed ideogram-install-auth setup
  • Familiarity with async/await and fetch API
  • Understanding of Ideogram response lifecycle (URLs expire)

Instructions

Step 1: Singleton Client with Auto-Download

// src/ideogram/client.ts
import { writeFileSync, mkdirSync } from "fs";
import { join } from "path";

const API_BASE = "https://api.ideogram.ai";

let instance: IdeogramClient | null = null;

export function getIdeogramClient(): IdeogramClient {
  if (!instance) {
    const key = process.env.IDEOGRAM_API_KEY;
    if (!key) throw new Error("IDEOGRAM_API_KEY not set");
    instance = new IdeogramClient(key);
  }
  return instance;
}

export class IdeogramClient {
  constructor(private apiKey: string) {}

  async generate(prompt: string, options: {
    model?: string;
    style_type?: string;
    aspect_ratio?: string;
    negative_prompt?: string;
    magic_prompt_option?: string;
    num_images?: number;
    seed?: number;
  } = {}) {
    const response = await fetch(`${API_BASE}/generate`, {
      method: "POST",
      headers: { "Api-Key": this.apiKey, "Content-Type": "application/json" },
      body: JSON.stringify({
        image_request: {
          prompt,
          model: options.model ?? "V_2",
          style_type: options.style_type ?? "AUTO",
          aspect_ratio: options.aspect_ratio ?? "ASPECT_1_1",
          magic_prompt_option: options.magic_prompt_option ?? "AUTO",
          negative_prompt: options.negative_prompt,
          num_images: options.num_images ?? 1,
          seed: options.seed,
        },
      }),
    });
    return this.handleResponse(response);
  }

  async describe(imagePath: string) {
    const form = new FormData();
    const file = new Blob([await import("fs").then(fs => fs.readFileSync(imagePath))]);
    form.append("image_file", file, "image.png");
    const response = await fetch(`${API_BASE}/describe`, {
      method: "POST",
      headers: { "Api-Key": this.apiKey },
      body: form,
    });
    return this.handleResponse(response);
  }

  async downloadImage(url: string, outputDir: string, filename?: string): Promise<string> {
    mkdirSync(outputDir, { recursive: true });
    const imgResponse = await fetch(url);
    if (!imgResponse.ok) throw new Error(`Download failed: ${imgResponse.status}`);
    const buffer = Buffer.from(await imgResponse.arrayBuffer());
    const outPath = join(outputDir, filename ?? `ideogram-${Date.now()}.png`);
    writeFileSync(outPath, buffer);
    return outPath;
  }

  private async handleResponse(response: Response) {
    if (!response.ok) {
      const body = await response.text();
      throw new IdeogramApiError(response.status, body);
    }
    return response.json();
  }
}

export class IdeogramApiError extends Error {
  constructor(public status: number, public body: string) {
    super(`Ideogram ${status}: ${body}`);
    this.name = "IdeogramApiError";
  }
  get isRateLimited() { return this.status === 429; }
  get isSafetyRejected() { return this.status === 422; }
  get isAuthError() { return this.status === 401; }
}

Step 2: Retry with Exponential Backoff

export async function withRetry<T>(
  operation: () => Promise<T>,
  config = { maxRetries: 3, baseDelayMs: 1000, maxDelayMs: 30000 }
): Promise<T> {
  for (let attempt = 0; attempt <= config.maxRetries; attempt++) {
    try {
      return await operation();
    } catch (err: any) {
      if (attempt === config.maxRetries) throw err;
      // Only retry on 429 (rate limit) or 5xx (server error)
      const status = err.status ?? err.response?.status;
      if (status && status !== 429 && status < 500) throw err;

      const delay = Math.min(
        config.baseDelayMs * Math.pow(2, attempt) + Math.random() * 500,
        config.maxDelayMs
      );
      console.warn(`Ideogram retry ${attempt + 1}/${config.maxRetries} in ${delay.toFixed(0)}ms`);
      await new Promise(r => setTimeout(r, delay));
    }
  }
  throw new Error("Unreachable");
}

// Usage: await withRetry(() => client.generate("a sunset"));

Step 3: Response Validation with Zod

import { z } from "zod";

const ImageResultSchema = z.object({
  url: z.string().url(),
  prompt: z.string(),
  resolution: z.string(),
  is_image_safe: z.boolean(),
  seed: z.number(),
  style_type: z.string().optional(),
});

const GenerateResponseSchema = z.object({
  created: z.string(),
  data: z.array(ImageResultSchema).min(1),
});

export function validateGenerateResponse(raw: unknown) {
  return GenerateResponseSchema.parse(raw);
}

Step 4: Python Client

# ideogram_client.py
import os, requests, hashlib, time
from pathlib import Path

class IdeogramClient:
    BASE_URL = "https://api.ideogram.ai"

    def __init__(self, api_key: str | None = None):
        self.api_key = api_key or os.environ.get("IDEOGRAM_API_KEY", "")
        if not self.api_key:
            raise ValueError("IDEOGRAM_API_KEY required")

    def generate(self, prompt: str, model="V_2", style_type="AUTO",
                 aspect_ratio="ASPECT_1_1", **kwargs) -> dict:
        resp = requests.post(f"{self.BASE_URL}/generate", headers=self._headers(),
            json={"image_request": {"prompt": prompt, "model": model,
                "style_type": style_type, "aspect_ratio": aspect_ratio,
                "magic_prompt_option": kwargs.get("magic_prompt", "AUTO"),
                **{k: v for k, v in kwargs.items() if k != "magic_prompt"}}})
        resp.raise_for_status()
        return resp.json()

    def download(self, url: str, output_dir: str = "./images") -> str:
        Path(output_dir).mkdir(parents=True, exist_ok=True)
        resp = requests.get(url)
        resp.raise_for_status()
        path = f"{output_dir}/ideogram-{int(time.time())}.png"
        Path(path).write_bytes(resp.content)
        return path

    def _headers(self):
        return {"Api-Key": self.api_key, "Content-Type": "application/json"}

Step 5: Multi-Tenant Factory

const tenantClients = new Map<string, IdeogramClient>();

export function getClientForTenant(tenantId: string): IdeogramClient {
  if (!tenantClients.has(tenantId)) {
    const apiKey = getTenantApiKey(tenantId); // from your secret store
    tenantClients.set(tenantId, new IdeogramClient(apiKey));
  }
  return tenantClients.get(tenantId)!;
}

Error Handling

PatternUse CaseBenefit
SingletonShared client across modulesSingle connection, consistent config
Retry wrapperRate limits and transient errorsAutomatic recovery from 429/5xx
Zod validationResponse validationCatches API changes at parse time
Auto-downloadImage persistencePrevents URL expiry data loss
Multi-tenantSaaS platformsPer-customer API key isolation

Output

  • Type-safe client singleton with generate and describe methods
  • Retry logic with exponential backoff and jitter
  • Runtime validation for API responses
  • Auto-download to prevent URL expiration issues

Resources

Next Steps

Apply patterns in ideogram-core-workflow-a for real-world usage.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 326,144. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.