
Wan 2.2 Turbo Infinite Image-to-Video API by Atlas Cloud
Image-to-video model for segmented prompt video generation with stable motion and 30fps workflow post-processing.
入力
出力
待機中各実行には$0.02かかります。$10で約500回実行できます。
次にできること:
コード例
import requests
import time
# Step 1: Start video generation
generate_url = "https://api.atlascloud.ai/api/v1/model/generateVideo"
headers = {
"Content-Type": "application/json",
"Authorization": "Bearer $ATLASCLOUD_API_KEY"
}
data = {
"model": "atlascloud/wan-2.2-turbo/infinite-image-to-video",
"prompt": "A beautiful sunset over the ocean with gentle waves",
"width": 512,
"height": 512,
"duration": 3,
"fps": 24,
}
generate_response = requests.post(generate_url, headers=headers, json=data)
generate_result = generate_response.json()
prediction_id = generate_result["data"]["id"]
# Step 2: Poll for result
poll_url = f"https://api.atlascloud.ai/api/v1/model/prediction/{prediction_id}"
def check_status():
while True:
response = requests.get(poll_url, headers={"Authorization": "Bearer $ATLASCLOUD_API_KEY"})
result = response.json()
if result["data"]["status"] in ["completed", "succeeded"]:
print("Generated video:", result["data"]["outputs"][0])
return result["data"]["outputs"][0]
elif result["data"]["status"] == "failed":
raise Exception(result["data"]["error"] or "Generation failed")
else:
# Still processing, wait 2 seconds
time.sleep(2)
video_url = check_status()インストール
お使いの言語に必要なパッケージをインストールしてください。
pip install requests認証
すべての API リクエストには API キーによる認証が必要です。API キーは Atlas Cloud ダッシュボードから取得できます。
export ATLASCLOUD_API_KEY="your-api-key-here"HTTP ヘッダー
import os
API_KEY = os.environ.get("ATLASCLOUD_API_KEY")
headers = {
"Content-Type": "application/json",
"Authorization": f"Bearer {API_KEY}"
}API キーをクライアントサイドのコードや公開リポジトリに公開しないでください。代わりに環境変数またはバックエンドプロキシを使用してください。
リクエストを送信
import requests
url = "https://api.atlascloud.ai/api/v1/model/generateVideo"
headers = {
"Content-Type": "application/json",
"Authorization": "Bearer $ATLASCLOUD_API_KEY"
}
data = {
"model": "your-model",
"prompt": "A beautiful landscape"
}
response = requests.post(url, headers=headers, json=data)
print(response.json())リクエストを送信
非同期生成リクエストを送信します。API は予測 ID を返し、それを使用してステータスの確認や結果の取得ができます。
/api/v1/model/generateVideoリクエストボディ
import requests
url = "https://api.atlascloud.ai/api/v1/model/generateVideo"
headers = {
"Content-Type": "application/json",
"Authorization": "Bearer $ATLASCLOUD_API_KEY"
}
data = {
"model": "atlascloud/wan-2.2-turbo/infinite-image-to-video",
"input": {
"prompt": "A beautiful sunset over the ocean with gentle waves"
}
}
response = requests.post(url, headers=headers, json=data)
result = response.json()
print(f"Prediction ID: {result['id']}")
print(f"Status: {result['status']}")レスポンス
{
"id": "pred_abc123",
"status": "processing",
"model": "model-name",
"created_at": "2025-01-01T00:00:00Z"
}ステータスを確認
予測エンドポイントをポーリングして、リクエストの現在のステータスを確認します。
/api/v1/model/prediction/{prediction_id}ポーリング例
import requests
import time
prediction_id = "pred_abc123"
url = f"https://api.atlascloud.ai/api/v1/model/prediction/{prediction_id}"
headers = { "Authorization": "Bearer $ATLASCLOUD_API_KEY" }
while True:
response = requests.get(url, headers=headers)
result = response.json()
status = result["data"]["status"]
print(f"Status: {status}")
if status in ["completed", "succeeded"]:
output_url = result["data"]["outputs"][0]
print(f"Output URL: {output_url}")
break
elif status == "failed":
print(f"Error: {result['data'].get('error', 'Unknown')}")
break
time.sleep(3)ステータス値
processingリクエストはまだ処理中です。completed生成が完了しました。出力が利用可能です。succeeded生成が成功しました。出力が利用可能です。failed生成に失敗しました。エラーフィールドを確認してください。完了レスポンス
{
"data": {
"id": "pred_abc123",
"status": "completed",
"outputs": [
"https://storage.atlascloud.ai/outputs/result.mp4"
],
"metrics": {
"predict_time": 45.2
},
"created_at": "2025-01-01T00:00:00Z",
"completed_at": "2025-01-01T00:00:10Z"
}
}ファイルをアップロード
Atlas Cloud ストレージにファイルをアップロードし、API リクエストで使用できる URL を取得します。multipart/form-data を使用してアップロードします。
/api/v1/model/uploadMediaアップロード例
import requests
url = "https://api.atlascloud.ai/api/v1/model/uploadMedia"
headers = { "Authorization": "Bearer $ATLASCLOUD_API_KEY" }
with open("image.png", "rb") as f:
files = {"file": ("image.png", f, "image/png")}
response = requests.post(url, headers=headers, files=files)
result = response.json()
download_url = result["data"]["download_url"]
print(f"File URL: {download_url}")レスポンス
{
"data": {
"download_url": "https://storage.atlascloud.ai/uploads/abc123/image.png",
"file_name": "image.png",
"content_type": "image/png",
"size": 1024000
}
}入力 Schema
以下のパラメータがリクエストボディで使用できます。
利用可能なパラメータはありません。
リクエストボディの例
{
"model": "atlascloud/wan-2.2-turbo/infinite-image-to-video"
}出力 Schema
API は生成された出力 URL を含む予測レスポンスを返します。
レスポンス例
{
"id": "pred_abc123",
"status": "completed",
"model": "model-name",
"outputs": [
"https://storage.atlascloud.ai/outputs/result.mp4"
],
"metrics": {
"predict_time": 45.2
},
"created_at": "2025-01-01T00:00:00Z",
"completed_at": "2025-01-01T00:00:10Z"
}Atlas Cloud Skills
Atlas Cloud Skills は 300 以上の AI モデルを AI コーディングアシスタントに直接統合します。ワンコマンドでインストールし、自然言語で画像・動画生成や LLM との対話が可能です。
対応クライアント
インストール
npx skills add AtlasCloudAI/atlas-cloud-skillsAPI キーの設定
Atlas Cloud ダッシュボードから API キーを取得し、環境変数として設定してください。
export ATLASCLOUD_API_KEY="your-api-key-here"機能
インストール後、AI アシスタントで自然言語を使用してすべての Atlas Cloud モデルにアクセスできます。
MCP Server
Atlas Cloud MCP Server は Model Context Protocol を通じて IDE と 300 以上の AI モデルを接続します。MCP 対応のあらゆるクライアントで動作します。
対応クライアント
インストール
npx -y atlascloud-mcp設定
以下の設定を IDE の MCP 設定ファイルに追加してください。
{
"mcpServers": {
"atlascloud": {
"command": "npx",
"args": [
"-y",
"atlascloud-mcp"
],
"env": {
"ATLASCLOUD_API_KEY": "your-api-key-here"
}
}
}
}利用可能なツール
APIスキーマ
スキーマが利用できませんWan 2.2 Turbo Infinite Image-to-Video
Model Overview
| Field | Description |
|---|---|
| Model Name | atlascloud/wan-2.2-turbo/infinite-image-to-video |
| Model Type | Advanced Image-to-Video Generation |
| Core Architecture | Mixture-of-Experts (MoE) |
| Active Parameters | 14B |
| Variant | Base |
Wan 2.2 Turbo Infinite Image-to-Video is an enhanced image-to-video model built on the Wan 2.2 foundation. Inheriting the Mixture-of-Experts (MoE) architecture and cinematic-level aesthetics of the original Wan series, this variant introduces two breakthroughs — inference acceleration and infinite-length generation — for unrestricted creative work.
Key Features & Innovations
1. Ultra-Fast Inference: 4-Step Distillation with RCM
To address the high latency typical of large-scale models, we apply specialized sampling optimization and knowledge distillation:
- RCM (Refined Consistency Model) Sampler — a more efficient ODE solver that significantly improves single-step sampling quality.
- 4-Step Distillation — denoising steps are compressed to 4 steps through multi-stage distillation, enabling cinematic-grade generation at a fraction of the original cost and unlocking low-latency interaction.
2. Infinite-Length Generation: Anchor-Frame Autoregressive Architecture
A targeted retraining gives the model an advanced temporal extension mechanism that breaks the duration limits of traditional video models:
- Anchor-Frame Evolution — automatically extracts key "anchor frames" during generation as global temporal references.
- Dual-Frame Constraint (Anchor + Last Frame) — combines structural consistency from the global anchor frame with motion continuity from the previous frame to construct video sequences autoregressively.
- Semantic Stability — subject identity, scene details, and lighting stay consistent across multi-minute outputs, suppressing semantic drift and logical collapse.
3. Cinematic-Level Aesthetics (Inherited)
The model retains the curated training foundation of Wan 2.2:
- Precise Control — detailed labels for lighting, composition, and color tone.
- Complex Motion — superior generation of realistic, fluid motion across diverse semantics.
Why Infinite?
Most image-to-video models lock you into a single short clip (5–10 s). Infinite extends that into a controlled multi-segment clip — output duration equals prompt_count × duration_per_segment, up to 6 prompts x 5 s. Direct each segment with its own prompt; the API returns one server-stitched 30 fps MP4.
| Prompts | Per-segment | Total output |
|---|---|---|
| 1 | 5 s | 5 s |
| 3 | 5 s | 15 s |
| 6 | 5 s | 30 s |
60-second Quickstart
curl -X POST https://api.atlascloud.ai/api/v1/model/generateVideo \ -H "Authorization: Bearer $APIKEY" \ -H "Content-Type: application/json" \ -d '{ "model": "atlascloud/wan-2.2-turbo/infinite-image-to-video", "image": "https://static.atlascloud.ai/media/images/db548fe3bd5cafa4ef7e0141d69c8566.jpeg", "prompt": [ "A classic golden Cadillac speeds through a desert, kicking up a massive cloud of dust behind it.", "Camera pans to the passenger firing an assault rifle at monstrous dinosaurs hot on the trail.", "The roaring creatures close in as the driver grips the wheel, knuckles white." ], "duration": 5, "resolution": "720p" }'
Returns one MP4 — segments are stitched server-side at 30 fps.
Base vs LoRA — which one?
| Base (this model) | LoRA variant | |
|---|---|---|
| Model name | atlascloud/wan-2.2-turbo/infinite-image-to-video | …/infinite-image-to-video-lora |
| Price (480 p, per second) | $0.020 | $0.026 (+30 %) |
| Best for | Standard runs, fast iteration, bulk drafts | Higher fidelity, fine-grained control |
| Recommended for | Pre-production, A/B prompts | Final renders |
Switch the variant by changing
modelonly — all other fields are identical.
Request Fields
| Field | Type | Required | Notes |
|---|---|---|---|
model | string | ✅ | atlascloud/wan-2.2-turbo/infinite-image-to-video |
image | string (URL) | ✅ | Source frame; jpg/png |
prompt | string[] | ✅ | Must be a JSON array. Plain string is rejected. |
duration | number | ✅ | Fixed at 5 s per segment. |
resolution | string | optional | 480p, 720p, or 1080p. Defaults to 720p. |
seed | number | optional | -1 for random |
Pricing — at a glance
price = $0.020 × max(1, prompt_count) × max(5, duration_seconds) × resolution_factor 480p → 1 720p → 2 1080p → 3
Common combos:
| Prompts | Duration | Resolution | Total |
|---|---|---|---|
| 1 | 5 s | 480 p | $0.10 |
| 1 | 5 s | 720 p | $0.20 |
| 1 | 5 s | 1080 p | $0.30 |
| 3 | 5 s | 720 p | $0.60 |
| 6 | 5 s | 720 p | $1.20 |
| 6 | 5 s | 1080 p | $1.80 |
Output Spec
- Format: MP4 (H.264)
- Frame rate: 30 fps (post-processed)
- Resolution: 480 p / 720 p / 1080 p tiers, aspect-ratio preserving
- Audio: none
Intended Use & Applications
- Cinematic Long-Take Production — high-fidelity, consistent long-duration shots without manual stitching.
- Low-Latency Interactive Content — leverage 4-step distillation for live broadcasts and AI-driven interactive installations.
- Advanced Image-to-Video (I2V) — transform a static image into infinite, naturally moving visual scrolls via anchor-frame technology.
- Professional Pre-visualization — minutes-long dynamic storyboards that compress pre-production time.
Usage Guidelines
This model is tuned for adult-oriented, unrestricted creative generation. By calling it you confirm:
- All depicted subjects are 18 +.
- You hold the rights to the source image.
- You will not generate content depicting real, identifiable people without their explicit consent.
Violations may result in account suspension.
Limitations
promptmust be a JSON array, never a plain string.- While anchor-frame technology suppresses cross-segment drift, it does not fully eliminate it — long prompts sharing fine identity details across many segments may still show minor variation.
- 480 p generates ~2× faster than 720 p; use 480 p for drafts.
Related
- LoRA variant:
atlascloud/wan-2.2-turbo/infinite-image-to-video-lora
Note: This model is designed to empower the creative community. Users are expected to follow AI ethical guidelines and copyright regulations.




