Documentation

Everything you need to start generating images and videos through one unified API.

Quickstart

Three steps to your first generation:

1. Create an account at panel.designapi.ink/register
2. Top up your balance and create an API key (Tokens → Add new)
3. Make your first request:

curl https://api.designapi.ink/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "flux",
    "prompt": "a futuristic city at sunset",
    "n": 1,
    "size": "1024x1024"
  }'

Authentication

All requests require an API key passed in the Authorization header:

Authorization: Bearer sk-...

Generate keys in your dashboard. You can create multiple keys, set per-key quotas, restrict allowed models, and revoke at any time.

Endpoints overview

Different models use different endpoint patterns. Pick the one matching your model:

Errors & rate limits

Errors follow OpenAI format:

{
  "error": {
    "message": "insufficient quota",
    "type": "one_api_error",
    "code": 403
  }
}

Image generation — DALL-E format

Works for: flux, flux-dev, flux-pro, flux-schnell, dall-e-3, recraftv3, ideogram-generate-v3, kling-image, doubao-seedream-*, bfl/flux-*, and most other static image models.

Request

POST https://api.designapi.ink/v1/images/generations
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

{
  "model": "flux-pro",
  "prompt": "a cyberpunk samurai under neon rain, ultra detailed",
  "n": 1,
  "size": "1024x1024",
  "response_format": "url"
}

Parameters

Python (OpenAI SDK)

from openai import OpenAI
client = OpenAI(api_key="YOUR_API_KEY", base_url="https://api.designapi.ink/v1")

result = client.images.generate(
    model="flux-pro",
    prompt="a cyberpunk samurai under neon rain",
    size="1024x1024",
    n=1,
)
print(result.data[0].url)

Node.js

import OpenAI from "openai";
const client = new OpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://api.designapi.ink/v1",
});
const r = await client.images.generate({
  model: "flux-pro",
  prompt: "a cyberpunk samurai under neon rain",
  size: "1024x1024",
});
console.log(r.data[0].url);

Image generation — chat-style

Some image models (Google Gemini, GPT-Image-via-chat, Sora image) return images through the chat completions endpoint. Works for: gemini-2.5-flash-image, gemini-3-pro-image-preview, nano-banana, nano-banana-pro, gpt-4o-image, sora_image.

curl https://api.designapi.ink/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nano-banana-pro",
    "messages": [
      {"role": "user", "content": "Generate an image of a futuristic city at sunset, 4K"}
    ]
  }'

The image URL or base64 will be returned inside choices[0].message.content as markdown ![](url) or as a direct base64 string.

Image editing

Models with the image-edit tag (Flux Kontext, SeedEdit, Gemini-image-edit, qwen-image-edit) accept an existing image plus an instruction.

Flux Kontext (DALL-E format with image)

curl https://api.designapi.ink/v1/images/edits \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "model=flux-kontext-pro" \
  -F "prompt=add sunglasses to the cat" \
  -F "image=@cat.png"

Gemini image edit (chat format)

{
  "model": "gemini-2.5-flash-image-preview",
  "messages": [{
    "role": "user",
    "content": [
      {"type": "text", "text": "Replace the background with a beach"},
      {"type": "image_url", "image_url": {"url": "https://..."}}
    ]
  }]
}

Midjourney — submit job

Midjourney is asynchronous. You submit a job, get a task_id, then poll until done.

1. Submit

POST https://api.designapi.ink/mj/submit/imagine
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

{
  "prompt": "a futuristic city at sunset --ar 16:9 --v 6",
  "botType": "MID_JOURNEY"
}

→ {"code": 1, "result": "1745324567890123", "description": "submitted"}

Use mj_fast_* models for fast queue, mj_relax_* for relax queue (slower but cheaper).

Midjourney — poll status

GET https://api.designapi.ink/mj/task/{task_id}/fetch
Authorization: Bearer YOUR_API_KEY

→ {
  "id": "1745324567890123",
  "status": "SUCCESS",
  "imageUrl": "https://cdn.../image.png",
  "buttons": [...]
}

Possible status values: NOT_START, SUBMITTED, IN_PROGRESS, SUCCESS, FAILURE.

Python polling helper

import requests, time
BASE, KEY = "https://api.designapi.ink", "YOUR_API_KEY"
H = {"Authorization": f"Bearer {KEY}"}

r = requests.post(f"{BASE}/mj/submit/imagine", headers=H, json={
    "prompt": "a futuristic city --ar 16:9", "botType": "MID_JOURNEY"
}).json()
task_id = r["result"]

while True:
    s = requests.get(f"{BASE}/mj/task/{task_id}/fetch", headers=H).json()
    if s["status"] == "SUCCESS":
        print(s["imageUrl"]); break
    if s["status"] == "FAILURE":
        raise RuntimeError(s.get("failReason"))
    time.sleep(3)

Midjourney — actions (upscale, variation, blend)

After a successful imagine, the response contains buttons with customId values. Use them to trigger actions:

POST /mj/submit/action
{
  "taskId": "PARENT_TASK_ID",
  "customId": "MJ::JOB::upsample::1::abc123"
}

Other endpoints: /mj/submit/blend (combine images), /mj/submit/describe (image to prompt), /mj/submit/modal (zoom/pan).

Video generation — async pattern

All video models are asynchronous. The flow is identical across providers:

1. Submit — POST returns task_id
2. Poll — GET task endpoint until status is completed or failed
3. Download — read the resulting video URL from the response

Generic example

POST https://api.designapi.ink/v1/video/generations
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

{
  "model": "kling-video-v2-6",
  "prompt": "a cinematic shot of a hummingbird",
  "duration": 5,
  "aspect_ratio": "16:9"
}

→ {"id": "vid_abc123", "status": "queued"}

GET https://api.designapi.ink/v1/video/generations/vid_abc123
Authorization: Bearer YOUR_API_KEY

→ {
  "id": "vid_abc123",
  "status": "completed",
  "video_url": "https://..."
}

Sora 2

Models: sora-2, sora-2-pro. Supports up to 15s (sora-2) or 25s (sora-2-pro).

{
  "model": "sora-2",
  "prompt": "a paper airplane gliding through a sunset sky, cinematic",
  "duration": 10,
  "size": "1280x720"
}

Sora-2 also supports the chat completions endpoint with messages instead of prompt — see the Sora API reference.

Kling

Models: kling-video-v1 through kling-video-v2-6, kling-video-v2-master (flagship), kling-effects, kling-lip-sync, kling-video-extend.

{
  "model": "kling-video-v2-6",
  "prompt": "a hummingbird hovering near a flower",
  "image_url": "https://...",   // optional, for i2v
  "duration": 5,
  "mode": "pro"                  // "std" or "pro"
}

Veo 3.1 (Google)

Models: veo3.1, veo3.1-fast, veo3.1-pro, veo3.1-pro-4k, veo3.1-components. Generates video with native audio.

{
  "model": "veo3.1-pro",
  "prompt": "ocean waves crashing on rocks at sunset, with sound of waves and seagulls",
  "duration": 8,
  "aspect_ratio": "16:9"
}

Runway Gen-3 / Gen-4

Models: runwayml-gen3a_turbo, runwayml-gen4_turbo, runway-act_one (face-to-character animation), runway-aleph (video editing).

{
  "model": "runwayml-gen4_turbo",
  "prompt": "a vintage car driving through Tokyo at night",
  "image_url": "https://...",   // for image-to-video
  "duration": 5
}

Billing & quota

• Each request deducts the model's price from your balance.
• Per-call models charge a fixed amount per request (regardless of output size).
• Token-based models charge per 1M input/output tokens.
• Top up via redeem codes — purchase from @claudxeseller.
• Detailed usage breakdown by model is available in your dashboard under Logs.

SDK compatibility

The base URL https://api.designapi.ink/v1 is fully OpenAI-compatible. You can use it as a drop-in replacement in any OpenAI SDK or library:

• Python: openai package — set base_url
• Node.js: openai package — set baseURL
• LangChain: set openai_api_base
• LiteLLM: set api_base
• Cursor / Continue.dev: custom OpenAI provider, base URL https://api.designapi.ink/v1

Need help? Contact @claudxeseller on Telegram.