Kikaku API

Quota Used

- / -

Remaining

Success Rate

Avg Response

Status: Email: Expires:

Generate Image

Model *

Prompt *

Reference Images (optional, max 14)

Size

Response Format

Sequential Image Gen

Max Images (1-15)

Prompt Optimization

Add watermark

Generated image will appear here

My Request Logs

Start Date

End Date

ID	Model	Status	Response Time	Timestamp	Error
Click "Filter" to load logs

📖 API Documentation

SeedDream 4.5 / 4.0 Image Generation API — Panduan Lengkap & Referensi

🧭 Overview

API ini adalah gateway ke model SeedDream untuk generate gambar berkualitas tinggi dari teks dan/atau gambar referensi. Didukung oleh model seedream-4-5-251128 dan seedream-4-0-250828.

🖼️ Multi-Mode Generation

Text-to-image, image-to-image, multi-image blending, dan batch generation dalam satu endpoint

📐 Flexible Resolution

Pilih preset (1K/2K/4K) atau custom pixel (WxH). Model otomatis menentukan aspect ratio sesuai prompt

⚡ Batch & Fallback

Generate hingga 15 gambar sekaligus. Jika model return kurang dari yang diminta, fallback otomatis dijalankan

🔑 Authentication

Setiap request harus menyertakan API key di header Authorization. Key berformat kikaku_XXXXXXX untuk AI Image, atau kikaku_s3_XXXXXXX untuk S3 Storage.

Authorization: Bearer kikaku_YOUR_API_KEY

⚠️ Penting: Jangan expose API key di frontend/client-side code. Selalu panggil API dari backend/server. Key yang bocor bisa di-revoke oleh admin kapan saja.

POST

/v1/images/generations

Endpoint utama untuk generate gambar. Mendukung text-to-image, image-to-image, multi-image blending, dan batch generation.

📋 Request Parameters

Parameter	Type	Required	Default	Penjelasan Detail
model	string	Required	—	Model yang digunakan untuk generate gambar. Pilihan: • `seedream-4.5` — Model terbaru, kualitas lebih tinggi, mendukung resolusi 2K & 4K. Mode prompt optimization hanya standard. • `seedream-4.0` — Model stabil, mendukung 1K/2K/4K. Mendukung mode fast untuk generate lebih cepat dengan trade-off kualitas.
prompt	string	Required	—	Deskripsi teks untuk gambar yang ingin di-generate. Maks 10.000 karakter, disarankan <600 kata bahasa Inggris. Tips: Gunakan bahasa natural yang koheren — deskripsikan subjek + aksi + environment. Jika estetika penting, tambahkan deskripsi style, warna, pencahayaan, atau komposisi. Contoh bagus: "Vibrant close-up editorial portrait, model with piercing gaze, wearing a sculptural hat, rich color blocking, sharp focus on eyes, shallow depth of field, Vogue magazine cover aesthetic" Peringatan: Prompt terlalu panjang bisa membuat model kehilangan fokus pada detail tertentu.
image	string \| string[]	Optional	—	Gambar referensi untuk image-to-image atau multi-image blending. Format input: • URL — URL yang bisa diakses publik, contoh: `https://example.com/photo.jpg` • Base64 — Format: `data:image/jpeg;base64,/9j/4AAQ...` Single image: Kirim sebagai string biasa. Multi-image (2–14): Kirim sebagai array of strings. Dalam prompt, referensikan dengan "image 1", "image 2", dst. Batasan input: JPEG, PNG, WEBP, BMP, TIFF, GIF — maks 10MB per gambar — aspect ratio [1/16, 16] — min 14×14 px — maks 6000×6000 total pixel (36 juta px).
size	string	Optional	2048x2048	Dimensi output gambar. Ada 2 metode (tidak bisa dipakai bersamaan): Metode 1 — Preset resolusi: Gunakan `2K` atau `4K` (4.5) / `1K`, `2K`, `4K` (4.0). Model otomatis menentukan width×height berdasarkan prompt. Deskripsikan aspect ratio di prompt, misal "poster portrait 3:4". Metode 2 — Custom pixel: Format `WIDTHxHEIGHT`, misal `2048x2048`. Harus memenuhi: total pixel dalam range & aspect ratio [1/16, 16].
response_format	string	Optional	"url"	Format response gambar yang di-generate: • `url` — Return URL download gambar. URL berlaku 24 jam setelah generate, pastikan download/simpan sebelum expired. • `b64_json` — Return data gambar sebagai Base64 string dalam JSON. Berguna ketika ingin langsung embed tanpa download.
watermark	boolean	Optional	true	Apakah menambahkan watermark ke gambar output. • `true` — Menambahkan teks "AI generated" di pojok kanan bawah gambar. • `false` — Tidak ada watermark. Cocok untuk produksi/commercial use.
sequential_image_generation	string	Optional	"disabled"	Kontrol fitur batch (generate beberapa gambar sekaligus). • `disabled` — Hanya generate 1 gambar (default). • `auto` — Model otomatis menentukan jumlah gambar berdasarkan isi prompt. Misal, "generate 4 seasonal scenes..." → model akan coba generate 4. ⚠️ Penting: Mode `auto` membuat model memutuskan sendiri berapa gambar yang dihasilkan. Jika prompt tidak eksplisit meminta multiple gambar, model mungkin hanya return 1. Sistem kami memiliki fallback otomatis: jika model return kurang dari `max_images`, request tambahan akan dibuat secara paralel.
sequential_image_generation_options.max_images	integer	Optional	15	Jumlah maksimum gambar yang di-generate dalam satu request batch. Range: 1–15. Constraint: Jumlah gambar referensi (input) + jumlah gambar output ≤ 15. Misal, jika kirim 3 gambar referensi, maka max_images efektif = 12. Hanya berlaku ketika `sequential_image_generation` = `"auto"`.
optimize_prompt_options.mode	string	Optional	"standard"	Mode optimisasi prompt sebelum generate dilakukan. • `standard` — Kualitas lebih tinggi, waktu generate lebih lama. Didukung semua model. • `fast` — Generate lebih cepat, tapi kualitas gambar rata-rata lebih rendah. Hanya untuk seedream-4.0. Use case fast mode: Preview/draft cepat, iterasi prompt, situasi di mana kecepatan lebih penting dari kualitas akhir.
n	integer	Legacy	1	Parameter legacy untuk backward-compatibility. Jumlah gambar yang ingin di-generate (1–15). Jika `n > 1` dan `sequential_image_generation` tidak diset, sistem otomatis mengkonversi ke: `sequential_image_generation: "auto"` + `max_images: n`. Rekomendasi: Gunakan `sequential_image_generation` + `max_images` secara eksplisit untuk kontrol penuh.

📦 Response Format

Semua response menggunakan format JSON. Output gambar selalu dalam format JPEG untuk kedua model (4.5 & 4.0).

// ✅ Success Response
{
  "model": "seedream-4-5-251128",
  "created": 1757323224,
  "data": [
    {
      "url": "https://ark-..../image-001.jpeg",   // jika response_format = "url"
      "size": "1760x2368"                          // dimensi aktual output
    },
    {
      "b64_json": "/9j/4AAQSkZJRgABAQ...",        // jika response_format = "b64_json"
      "size": "2048x2048"
    }
  ],
  "usage": {
    "generated_images": 1,                         // jumlah gambar yang berhasil di-generate
    "output_tokens": 16280,                        // token cost (width*height/256)
    "total_tokens": 16280
  }
}

// ❌ Error Response (top-level)
{
  "error": {
    "code": "InvalidParameter",
    "message": "Invalid size \"5K\". Use 2K/4K or WIDTHxHEIGHT"
  }
}

// ⚠️ Partial Error (dalam batch — satu gambar gagal, lainnya berhasil)
{
  "data": [
    { "url": "https://...", "size": "2048x2048" },
    { "error": { "code": "ContentFilter", "message": "Image rejected by content filter" } }
  ]
}

💡 Catatan Response:

data adalah array — bahkan untuk single image, tetap array dengan 1 elemen.
data[].size berformat "WxH" pixel (misal "2048x2048"), menunjukkan dimensi aktual gambar.
Pada batch mode, jika satu gambar gagal karena content filter, gambar lainnya tetap dikembalikan.
Jika gagal karena internal error (500), generate berikutnya dalam batch dibatalkan.
usage.output_tokens dihitung per gambar: ⌈(width × height) / 256⌉. Billing berdasarkan gambar yang berhasil saja.

🎯 Use Cases & Contoh

1️⃣

Text-to-Image — Generate gambar dari teks saja

Berikan instruksi teks yang jelas dan akurat. Model akan generate gambar berkualitas tinggi yang sesuai deskripsi. Tidak perlu parameter image.

curl -X POST YOUR_BASE_URL/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer kikaku_YOUR_KEY" \
  -d '{
    "model": "seedream-4.5",
    "prompt": "Vibrant close-up editorial portrait, model with piercing gaze, wearing a sculptural hat, rich color blocking, sharp focus on eyes, shallow depth of field, Vogue magazine cover aesthetic, dramatic studio lighting",
    "size": "2K",
    "response_format": "url",
    "watermark": false
}'

2️⃣

Image-to-Image — Edit gambar dengan instruksi teks

Kirim 1 gambar referensi + instruksi perubahan melalui prompt. Cocok untuk menambah/hapus elemen, ubah style/tekstur, ganti background, adjust warna, atau ubah perspektif.

curl -X POST YOUR_BASE_URL/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer kikaku_YOUR_KEY" \
  -d '{
    "model": "seedream-4.5",
    "prompt": "Keep the model pose unchanged. Change the clothing material from silver metal to completely transparent clear water. Through the liquid, the skin details are visible. Lighting changes from reflection to refraction.",
    "image": "https://example.com/my-photo.jpg",
    "size": "2K",
    "watermark": false
}'

3️⃣

Multi-Image Blending — Gabungkan elemen dari beberapa gambar

Kirim 2–14 gambar referensi sebagai array. Referensikan dengan "image 1", "image 2", dst. dalam prompt. Contoh use case: combine baju dari foto A ke model di foto B, merge style dari beberapa gambar, virtual try-on.

curl -X POST YOUR_BASE_URL/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer kikaku_YOUR_KEY" \
  -d '{
    "model": "seedream-4.5",
    "prompt": "Replace the clothing in image 1 with the outfit from image 2. Keep the original pose and background.",
    "image": [
      "https://example.com/model-photo.jpg",
      "https://example.com/outfit-photo.jpg"
    ],
    "size": "2K",
    "sequential_image_generation": "disabled",
    "watermark": false
}'

4️⃣

Text-to-Batch — Generate set gambar terkait dari teks

Generate kumpulan gambar tematik — misal storyboard film, visual branding, variasi seasonal. Set sequential_image_generation: "auto" dan tentukan max_images.
Tips: Sebutkan jumlah gambar dan deskripsi masing-masing secara eksplisit dalam prompt agar model tahu berapa gambar yang harus di-generate.

curl -X POST YOUR_BASE_URL/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer kikaku_YOUR_KEY" \
  -d '{
    "model": "seedream-4.5",
    "prompt": "Generate 4 illustrations of the same courtyard across seasons: Scene 1 spring with cherry blossoms, Scene 2 summer with bright sunlight, Scene 3 autumn with falling leaves, Scene 4 winter with snow covering.",
    "size": "2K",
    "sequential_image_generation": "auto",
    "sequential_image_generation_options": { "max_images": 4 },
    "watermark": false
}'

5️⃣

Image-to-Batch — Generate batch dari 1 gambar referensi

Kirim 1 gambar referensi + prompt → generate beberapa gambar terkait. Misal: dari 1 logo generate visual system (packaging, hat, card, lanyard).

curl -X POST YOUR_BASE_URL/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer kikaku_YOUR_KEY" \
  -d '{
    "model": "seedream-4.5",
    "prompt": "Using this LOGO as reference, create a visual design system for an outdoor sports brand named GREEN, including packaging bags, hats, cards, lanyards. Main tone is green, fun, simple, modern style.",
    "image": "https://example.com/my-logo.png",
    "size": "2K",
    "sequential_image_generation": "auto",
    "sequential_image_generation_options": { "max_images": 4 },
    "watermark": false
}'

6️⃣

Multi-Image-to-Batch — Batch dari beberapa gambar referensi

Kirim 2+ gambar referensi + batch mode → generate set gambar yang konsisten dengan kedua referensi. Misal: karakter + objek → generate variasi scene.

curl -X POST YOUR_BASE_URL/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer kikaku_YOUR_KEY" \
  -d '{
    "model": "seedream-4.5",
    "prompt": "Generate 3 images of a girl and a cow plushie happily riding a roller coaster, depicting morning, noon, and night scenes.",
    "image": [
      "https://example.com/girl-character.png",
      "https://example.com/cow-plushie.png"
    ],
    "sequential_image_generation": "auto",
    "sequential_image_generation_options": { "max_images": 3 },
    "size": "2K",
    "watermark": false
}'

7️⃣

Fast Mode — Generate cepat (seedream-4.0 only)

Gunakan optimize_prompt_options.mode: "fast" untuk generate lebih cepat. Trade-off: kualitas gambar sedikit lebih rendah dari mode standard. Hanya tersedia di seedream-4.0.

curl -X POST YOUR_BASE_URL/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer kikaku_YOUR_KEY" \
  -d '{
    "model": "seedream-4.0",
    "prompt": "Generate 4 coherent illustrations of a courtyard across four seasons, unified style capturing unique colors and atmosphere of each season.",
    "size": "2K",
    "sequential_image_generation": "auto",
    "sequential_image_generation_options": { "max_images": 4 },
    "optimize_prompt_options": { "mode": "fast" },
    "watermark": false
}'

8️⃣

Legacy `n` Parameter — Backward Compatibility

Jika masih menggunakan parameter n, sistem otomatis mengkonversi ke batch mode. Tapi disarankan gunakan sequential_image_generation secara eksplisit.

curl -X POST YOUR_BASE_URL/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer kikaku_YOUR_KEY" \
  -d '{
    "model": "seedream-4.5",
    "prompt": "A cute cat in different poses: sitting, jumping, sleeping",
    "n": 3,
    "size": "2K",
    "watermark": false
}'
// ↑ Secara internal dikonversi menjadi:
// sequential_image_generation: "auto", max_images: 3

📐 Dimensi Output Gambar — Panduan Detail

Dua metode sizing tersedia. Jangan gunakan keduanya bersamaan.

Metode 1 — Preset Resolusi

Pilih resolusi (1K/2K/4K), lalu deskripsikan aspect ratio/bentuk/tujuan di prompt. Model otomatis menentukan width×height.

"size": "2K"
"prompt": "...a tall portrait poster 3:4..."

Metode 2 — Custom Pixel

Tentukan width×height eksak dalam pixel. Harus memenuhi total pixel range DAN aspect ratio [1/16, 16].

"size": "2848x1600" // 16:9 landscape
"size": "1600x2848" // 9:16 portrait

seedream-4.5 — Recommended Dimensions

Resolution	1:1	4:3	3:4	16:9	9:16	3:2	2:3	21:9
2K	2048×2048	2304×1728	1728×2304	2848×1600	1600×2848	2496×1664	1664×2496	3136×1344
4K	4096×4096	4704×3520	3520×4704	5504×3040	3040×5504	4992×3328	3328×4992	6240×2656

Pixel range: [2560×1440 = 3,686,400 , 4096×4096 = 16,777,216]

seedream-4.0 — Recommended Dimensions

Resolution	1:1	4:3	3:4	16:9	9:16	3:2	2:3	21:9
1K	1024×1024	1152×864	864×1152	1312×736	736×1312	1248×832	832×1248	1568×672
2K	2048×2048	2304×1728	1728×2304	2848×1600	1600×2848	2496×1664	1664×2496	3136×1344
4K	4096×4096	4704×3520	3520×4704	5504×3040	3040×5504	4992×3328	3328×4992	6240×2656

Pixel range: [1280×720 = 921,600 , 4096×4096 = 16,777,216]

💡 Validasi Custom Pixel: Kedua syarat harus terpenuhi: (1) Total pixel (W×H) dalam range model, DAN (2) Aspect ratio (W/H) antara 1/16 s.d. 16.
✅ Contoh valid (4.5): 3750×1250 → total = 4,687,500 (dalam range), ratio = 3 (OK)
❌ Contoh invalid (4.5): 1500×1500 → total = 2,250,000 (di bawah minimum 3,686,400)

📋 Perbandingan Model

Tabel lengkap perbandingan fitur antara kedua model yang didukung.

Fitur	seedream-4.5	seedream-4.0
Model ID (internal)	seedream-4-5-251128	seedream-4-0-250828
Text to Image	✅ Didukung	✅ Didukung
Single Image → Image	✅ Didukung	✅ Didukung
Multi-Image Blending	✅ Hingga 14 referensi	✅ Hingga 14 referensi
Text → Batch Images	✅ Hingga 15 output	✅ Hingga 15 output
Image → Batch Images	✅ Input+output ≤ 15	✅ Input+output ≤ 15
Multi-Image → Batch Images	✅ Input+output ≤ 15	✅ Input+output ≤ 15
Resolusi yang Didukung	2K 4K	1K 2K 4K
Output Format	JPEG only (tidak bisa custom)	JPEG only (tidak bisa custom)
Prompt Optimization	standard	standard fast
Max Gambar per Menit	500	500
Max Referensi Input	14 gambar	14 gambar
Watermark	✅ Bisa on/off	✅ Bisa on/off

🔗 Endpoint Lainnya

GET /v1/usage — Cek quota & statistik penggunaan

Menampilkan informasi quota, jumlah pemakaian, sisa quota, status key, dan statistik request (total, sukses, gagal, rata-rata response time).

curl -H "Authorization: Bearer kikaku_YOUR_KEY" YOUR_BASE_URL/v1/usage

// Response:
{
  "owner_name": "John Doe",
  "email": "[email protected]",
  "quota_limit": 1000,
  "used_count": 42,
  "remaining": 958,
  "status": "active",
  "created_at": "2025-01-15T10:00:00Z",
  "expires_at": "2026-01-15T10:00:00Z",
  "stats": {
    "total_requests": 50,
    "successful_requests": 42,
    "failed_requests": 8,
    "avg_response_time_ms": 3200
  }
}

GET /v1/logs — Riwayat request log

Query parameters:

Param	Type	Keterangan
start_date	string	Filter mulai tanggal (ISO format), misal: 2025-01-01
end_date	string	Filter sampai tanggal
page	integer	Halaman (default: 1)
limit	integer	Jumlah per halaman (default: 20, maks: 100)

curl -H "Authorization: Bearer kikaku_YOUR_KEY" \
  "YOUR_BASE_URL/v1/logs?start_date=2025-01-01&end_date=2025-12-31&page=1&limit=20"

// Response:
{
  "logs": [
    {
      "id": 123,
      "model": "seedream-4.5",
      "timestamp": "2025-06-15T14:30:00Z",
      "status": "success",
      "response_time_ms": 2800,
      "error_message": null
    }
  ],
  "pagination": { "page": 1, "limit": 20, "total": 150 }
}

GET /v1/gallery — Galeri gambar yang pernah di-generate

Semua gambar yang berhasil di-generate otomatis disimpan ke S3 storage dan bisa di-browse lewat endpoint ini. Query parameters:

Param	Type	Keterangan
page	integer	Halaman (default: 1)
limit	integer	Jumlah per halaman (default: 20, maks: 100)

curl -H "Authorization: Bearer kikaku_YOUR_KEY" \
  "YOUR_BASE_URL/v1/gallery?page=1&limit=20"

// Response:
{
  "images": [
    {
      "id": 456,
      "prompt": "A magical forest with glowing mushrooms",
      "model": "seedream-4.5",
      "s3_url": "https://s3.onidel.com/bucket/path/image.jpeg",
      "image_size": "2K",
      "created_at": "2025-06-15T14:30:05Z"
    }
  ],
  "pagination": { "page": 1, "limit": 20, "total": 42 }
}

⚠️ Batasan Input Gambar

Batasan	Nilai	Detail
Format yang didukung	JPEG, PNG, WEBP, BMP, TIFF, GIF	Semua format umum. GIF diambil frame pertama.
Ukuran file	Maks 10 MB per gambar	Kompresi dulu jika melebihi limit.
Dimensi minimum	14 × 14 pixel	Width dan height harus > 14px.
Total pixel maksimum	6000 × 6000 = 36 juta px	Dihitung dari W×H (bukan per dimensi).
Aspect ratio	[1/16, 16]	Ratio W/H harus antara 1/16 s.d. 16.
Jumlah gambar referensi	Maks 14 gambar	Untuk multi-image blending / batch generation.
Input + Output total	≤ 15	Jumlah gambar referensi + jumlah gambar output tidak boleh melebihi 15.
Base64 format	data:image/jpeg;base64,...	Prefix wajib, format gambar lowercase.

⚡ Rate Limits, Errors & Troubleshooting

HTTP Status Codes

Code	Arti	Cara Mengatasi
`200`	Success	Request berhasil. Gambar tersedia di `data[]`.
`400`	Bad Request	Parameter tidak valid. Periksa model, size, format image, dan semua field yang dikirim.
`401`	Unauthorized	API key tidak valid atau tidak ada. Pastikan format header benar: `Bearer kikaku_xxx`.
`403`	Forbidden	Key suspended, expired, atau quota habis. Hubungi admin untuk perpanjangan.
`429`	Rate Limit	Terlalu banyak request. Tunggu sebentar lalu coba lagi. Rate limit: 10 req/menit per key.
`500`	Server Error	Error internal. Coba lagi. Jika terus terjadi, hubungi support.
`timeout`	Request Timeout	Request melebihi 120 detik. Coba kurangi resolusi atau jumlah batch.

Batasan & Catatan Penting

🕐 Rate Limits

10 requests/menit per API key
Maks 500 gambar/menit per model (RPM server-side)
Batch request dengan fallback dihitung sebagai multiple request pada quota

💾 Penyimpanan & Expiry

URL gambar dari BytePlus expires 24 jam — simpan segera
Gambar otomatis di-backup ke S3 storage setelah generate
Gallery bisa diakses via GET /v1/gallery

🔄 Batch Fallback

Mode auto membuat model sendiri yang tentukan jumlah output
Jika kurang dari max_images, sistem buat request tambahan paralel
Setiap request tambahan mengurangi quota sebagai 1 API call

📝 Prompt Tips

Gunakan <600 kata Inggris, bahasa natural & koheren
Deskripsikan: subjek + aksi + environment + style
Untuk batch: sebutkan jumlah & deskripsi tiap gambar eksplisit
Untuk sizing: deskripsikan aspect ratio di prompt jika pakai preset resolusi

Storage Used

- / -

Remaining

Status

Bucket: - Email: -

Upload File

File

Path / Key (optional, defaults to filename)

Files

Name	Size	Last Modified	Actions
Loading files...

🔑 S3-Compatible Credentials

Use these credentials with any S3-compatible tool (AWS CLI, boto3, aws-sdk, etc.)

Endpoint URL

-

Bucket Name

-

Access Key

-

Secret Key

-

Region

-

Quick Start Examples

AWS CLI

# Configure
aws configure set aws_access_key_id YOUR_ACCESS_KEY
aws configure set aws_secret_access_key YOUR_SECRET_KEY
aws configure set default.region YOUR_REGION

# List files
aws s3 ls s3://YOUR_BUCKET --endpoint-url YOUR_ENDPOINT

# Upload file
aws s3 cp myfile.txt s3://YOUR_BUCKET/myfile.txt --endpoint-url YOUR_ENDPOINT

# Download file
aws s3 cp s3://YOUR_BUCKET/myfile.txt ./myfile.txt --endpoint-url YOUR_ENDPOINT

Python (boto3)

import boto3

s3 = boto3.client('s3',
    endpoint_url='YOUR_ENDPOINT',
    aws_access_key_id='YOUR_ACCESS_KEY',
    aws_secret_access_key='YOUR_SECRET_KEY',
    region_name='YOUR_REGION')

# Upload
s3.upload_file('myfile.txt', 'YOUR_BUCKET', 'myfile.txt')

# Download
s3.download_file('YOUR_BUCKET', 'myfile.txt', './myfile.txt')

# List
response = s3.list_objects_v2(Bucket='YOUR_BUCKET')
for obj in response.get('Contents', []):
    print(obj['Key'], obj['Size'])

Node.js (aws-sdk v3)

const { S3Client, PutObjectCommand, ListObjectsV2Command } = require('@aws-sdk/client-s3');
const fs = require('fs');

const s3 = new S3Client({
  endpoint: 'YOUR_ENDPOINT',
  region: 'YOUR_REGION',
  credentials: {
    accessKeyId: 'YOUR_ACCESS_KEY',
    secretAccessKey: 'YOUR_SECRET_KEY',
  },
  forcePathStyle: true,
});

// Upload
await s3.send(new PutObjectCommand({
  Bucket: 'YOUR_BUCKET',
  Key: 'myfile.txt',
  Body: fs.readFileSync('myfile.txt'),
}));

// List
const { Contents } = await s3.send(new ListObjectsV2Command({
  Bucket: 'YOUR_BUCKET',
}));
Contents.forEach(obj => console.log(obj.Key, obj.Size));

📖 S3 Storage API Documentation

All requests use your API key for authentication. Files are proxied through our server with quota enforcement.

🔑 Authentication

Include your API key in every request:

Authorization: Bearer kikaku_s3_YOUR_API_KEY

POST /s3/upload

Upload a file. Use multipart/form-data with field file. Optional key field for custom path.

curl -X POST YOUR_BASE_URL/s3/upload \
  -H "Authorization: Bearer YOUR_S3_KEY" \
  -F "[email protected]" \
  -F "key=folder/myfile.txt"

GET /s3/files

List files in your bucket. Optional query params: prefix, limit, cursor.

curl -H "Authorization: Bearer YOUR_S3_KEY" \
  YOUR_BASE_URL/s3/files?prefix=folder/&limit=50

GET /s3/files/{key}

Download a file by its key.

curl -H "Authorization: Bearer YOUR_S3_KEY" \
  YOUR_BASE_URL/s3/files/folder/myfile.txt -o myfile.txt

DELETE /s3/files/{key}

Delete a file. Freed storage is reclaimed from your quota.

curl -X DELETE -H "Authorization: Bearer YOUR_S3_KEY" \
  YOUR_BASE_URL/s3/files/folder/myfile.txt

GET /s3/usage

Check your storage quota and usage.

curl -H "Authorization: Bearer YOUR_S3_KEY" \
  YOUR_BASE_URL/s3/usage

⚡ Rate Limits & Errors

60 requests/minute per API key
500 MB max file size per upload
401 — Invalid/missing API key
403 — Key suspended, expired, or quota exceeded
404 — File not found
429 — Rate limit exceeded

📝 Node.js Example

const fs = require('fs');
const FormData = require('form-data');
const axios = require('axios');

const API_KEY = 'YOUR_S3_KEY';
const BASE = 'YOUR_BASE_URL';

// Upload
const form = new FormData();
form.append('file', fs.createReadStream('photo.jpg'));
form.append('key', 'images/photo.jpg');
await axios.post(`${BASE}/s3/upload`, form, {
  headers: { ...form.getHeaders(), Authorization: `Bearer ${API_KEY}` }
});

// List files
const { data } = await axios.get(`${BASE}/s3/files`, {
  headers: { Authorization: `Bearer ${API_KEY}` }
});
console.log(data.files);

// Download
const resp = await axios.get(`${BASE}/s3/files/images/photo.jpg`, {
  headers: { Authorization: `Bearer ${API_KEY}` },
  responseType: 'stream'
});
resp.data.pipe(fs.createWriteStream('downloaded.jpg'));

🐍 Python Example

import requests

API_KEY = 'YOUR_S3_KEY'
BASE = 'YOUR_BASE_URL'
headers = {'Authorization': f'Bearer {API_KEY}'}

# Upload
with open('photo.jpg', 'rb') as f:
    resp = requests.post(f'{BASE}/s3/upload',
        headers=headers,
        files={'file': f},
        data={'key': 'images/photo.jpg'})
print(resp.json())

# List files
resp = requests.get(f'{BASE}/s3/files', headers=headers)
for f in resp.json()['files']:
    print(f['key'], f['size'])

# Download
resp = requests.get(f'{BASE}/s3/files/images/photo.jpg',
    headers=headers, stream=True)
with open('downloaded.jpg', 'wb') as f:
    for chunk in resp.iter_content(8192):
        f.write(chunk)

No Active Features

AI Image Generator

Generate Image

🎨 My Gallery

My Request Logs

📖 API Documentation

🧭 Overview

🔑 Authentication

/v1/images/generations

📋 Request Parameters

📦 Response Format

🎯 Use Cases & Contoh

Text-to-Image — Generate gambar dari teks saja

Image-to-Image — Edit gambar dengan instruksi teks

Multi-Image Blending — Gabungkan elemen dari beberapa gambar

Text-to-Batch — Generate set gambar terkait dari teks

Image-to-Batch — Generate batch dari 1 gambar referensi

Multi-Image-to-Batch — Batch dari beberapa gambar referensi

Fast Mode — Generate cepat (seedream-4.0 only)

Legacy n Parameter — Backward Compatibility

📐 Dimensi Output Gambar — Panduan Detail

Metode 1 — Preset Resolusi

Metode 2 — Custom Pixel

seedream-4.5 — Recommended Dimensions

seedream-4.0 — Recommended Dimensions

📋 Perbandingan Model

🔗 Endpoint Lainnya

⚠️ Batasan Input Gambar

⚡ Rate Limits, Errors & Troubleshooting

HTTP Status Codes

Batasan & Catatan Penting

S3 Storage

Upload File

Files

🔑 S3-Compatible Credentials

Quick Start Examples

AWS CLI

Python (boto3)

Node.js (aws-sdk v3)

📖 S3 Storage API Documentation

🔑 Authentication

POST /s3/upload

GET /s3/files

GET /s3/files/{key}

DELETE /s3/files/{key}

GET /s3/usage

⚡ Rate Limits & Errors

📝 Node.js Example

🐍 Python Example

Your API Key Details

Edit Profile

Legacy `n` Parameter — Backward Compatibility