Generate your first image - YingTu API Docs

This walkthrough uses gemini-3.1-flash-image at 1K. It is a practical default for a new integration because the same model can later move to 2K or 4K without changing the request shape. Before you begin, create a development key in the YingTu console and review authentication safety. Run the code on a server or local development machine, not in browser-delivered application code.

Send the request

cURL

curl --request POST \
  --url 'https://api.yingtu.ai/v1beta/models/gemini-3.1-flash-image:generateContent' \
  --header 'Authorization: Bearer YOUR_YINGTU_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "contents": [{
      "role": "user",
      "parts": [{"text": "A clear editorial illustration of a small solar-powered library"}]
    }],
    "generationConfig": {
      "responseModalities": ["TEXT", "IMAGE"],
      "imageConfig": {"aspectRatio": "1:1", "imageSize": "1K"},
      "candidateCount": 1
    }
  }'

The request is billable. Keep candidateCount omitted or set to 1; send separate requests when you need separate images.

Save the image in Python

import base64
import os
from pathlib import Path

import requests

response = requests.post(
    "https://api.yingtu.ai/v1beta/models/"
    "gemini-3.1-flash-image:generateContent",
    headers={"Authorization": f"Bearer {os.environ['YINGTU_API_KEY']}"},
    json={
        "contents": [{
            "role": "user",
            "parts": [{
                "text": "A clear editorial illustration of a small solar-powered library"
            }],
        }],
        "generationConfig": {
            "responseModalities": ["TEXT", "IMAGE"],
            "imageConfig": {"aspectRatio": "1:1", "imageSize": "1K"},
            "candidateCount": 1,
        },
    },
    timeout=180,
)
response.raise_for_status()
payload = response.json()

image_part = next(
    (
        part
        for candidate in payload.get("candidates", [])
        for part in candidate.get("content", {}).get("parts", [])
        if part.get("inlineData", {}).get("data")
    ),
    None,
)

if image_part is None:
    feedback = payload.get("promptFeedback", {})
    reason = feedback.get("blockReason", "no image part returned")
    raise RuntimeError(f"Generation completed without an image: {reason}")

inline = image_part["inlineData"]
extensions = {
    "image/png": ".png",
    "image/jpeg": ".jpg",
    "image/webp": ".webp",
}
extension = extensions.get(inline["mimeType"])
if extension is None:
    raise RuntimeError(f"Unsupported returned MIME: {inline['mimeType']}")

Path(f"yingtu-first-image{extension}").write_bytes(
    base64.b64decode(inline["data"])
)

Do not assume the image is parts[0] or that it is always PNG. The response may use PNG, JPEG, or WebP and may place a text part before the image; always scan the complete parts array and use the returned MIME type.

The same parser in Node.js

import { writeFile } from "node:fs/promises";

const response = await fetch(
  "https://api.yingtu.ai/v1beta/models/" +
    "gemini-3.1-flash-image:generateContent",
  {
    method: "POST",
    headers: {
      "Authorization": `Bearer ${process.env.YINGTU_API_KEY}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      contents: [{
        role: "user",
        parts: [{ text: "A compact cut-paper map of a coastal town" }],
      }],
      generationConfig: {
        responseModalities: ["TEXT", "IMAGE"],
        imageConfig: { aspectRatio: "1:1", imageSize: "1K" },
        candidateCount: 1,
      },
    }),
  },
);

if (!response.ok) {
  throw new Error(`Generation failed with HTTP ${response.status}`);
}

const payload = await response.json();
const parts = payload.candidates?.flatMap(
  (candidate) => candidate.content?.parts ?? [],
) ?? [];
const imagePart = parts.find((part) => part.inlineData?.data);

if (!imagePart) {
  throw new Error(
    `No image returned: ${payload.promptFeedback?.blockReason ?? "unknown reason"}`,
  );
}

const mime = imagePart.inlineData.mimeType;
const extension = {
  "image/png": ".png",
  "image/jpeg": ".jpg",
  "image/webp": ".webp",
}[mime];

if (!extension) throw new Error(`Unsupported returned MIME: ${mime}`);
await writeFile(
  `yingtu-first-image${extension}`,
  Buffer.from(imagePart.inlineData.data, "base64"),
);

Choose another model

Compare the current 1K, multi-resolution, premium, and migration paths.

Move work to a task

Submit the same body and retrieve the result after the connection ends.

Diagnose failures

Distinguish validation, authentication, balance, rate-limit, and server errors.

Check SDK support

Use the maintained REST path without assuming an SDK is compatible.

​Send the request

​Save the image in Python

​The same parser in Node.js