InfiniteTalk turns a reference portrait and audio into a realistic talking-head video with lip-sync, supporting up to 10-minute audio in 480p or 720p.
入力
出力
待機中各実行には$0.03かかります。$10で約333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/infinitetalk",
"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/infinitetalk",
"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/infinitetalk"
}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InfiniteTalk: Audio-Driven Talking Video Generation
1. Introduction
InfiniteTalk is an audio-driven video generation model developed by AtlasCloud that transforms a single portrait image into a realistic talking-head video synchronized to any speech audio input. Built on a modified Wan2.1 I2V-14B diffusion transformer backbone with a dedicated audio cross-attention module, InfiniteTalk achieves phoneme-level lip synchronization while preserving the subject's identity, hairstyle, clothing, and background throughout the entire video.
InfiniteTalk's core innovation lies in its triple cross-attention architecture: each transformer block processes visual self-attention, text prompt cross-attention, and frame-level audio cross-attention in sequence, enabling precise per-frame audio-visual alignment. Combined with a streaming inference pipeline that processes video in overlapping segments, InfiniteTalk supports continuous video generation of up to 10 minutes from a single request — far exceeding the typical 5–15 second limit of conventional image-to-video models. The model also supports dual-person mode, animating two speakers simultaneously within the same frame using separate audio tracks and bounding box annotations.
2. Key Features & Innovations
Triple Cross-Attention Audio Conditioning: Unlike text-only conditioned video models, InfiniteTalk injects audio embeddings at every transformer block via a dedicated cross-attention layer. Audio features are extracted frame-by-frame using a Wav2Vec2 encoder, providing per-frame speech signal anchoring that drives natural mouth movements, facial micro-expressions, and head motion synchronized to the audio input.
Streaming Long-Form Video Generation: InfiniteTalk's streaming mode processes audio in overlapping clip segments with configurable motion frame overlap, automatically concatenating segments into seamless long-form video. This enables generation of minutes-long talking videos without quality degradation or identity drift — a capability not available in standard image-to-video pipelines limited to single-shot outputs.
High-Fidelity Identity Preservation: The model maintains consistent facial identity, hairstyle, clothing texture, and background composition across the entire generated video. The audio conditioning signal provides strong per-frame constraints that prevent the identity drift commonly observed in long unconditional video generation.
Dual-Person Conversation Mode: InfiniteTalk supports animating two speakers in a single scene by accepting separate audio tracks and bounding box coordinates for each person. This enables realistic conversation scenarios, interview formats, and dialogue-driven content without requiring separate generation passes or post-production compositing.
Flexible Input Modalities: The model accepts either a static portrait image or a reference video as the visual source, combined with audio in WAV or MP3 format. Text prompts provide additional guidance for expression style, posture, and behavioral nuance, giving creators fine-grained control over the generated output.
Conditional VSR Upscaling: When generating at 720p resolution with audio duration under 60 seconds, InfiniteTalk automatically routes output through a FlashVSR super-resolution pipeline, delivering enhanced visual clarity without additional user configuration or cost management.
3. Model Architecture & Technical Details
InfiniteTalk is built on the Wan2.1 I2V-14B foundation model (14 billion parameters, 480p native resolution) with custom InfiniteTalk adapter weights that introduce the audio cross-attention pathway. The audio encoder uses a Chinese-Wav2Vec2-Base model that extracts frame-aligned speech embeddings at 25 fps video rate, creating a one-to-one correspondence between audio features and generated video frames.
The inference pipeline operates in two modes. In clip mode, the model generates a single video segment of up to 81 frames (approximately 3.2 seconds at 25 fps), suitable for short-form content. In streaming mode, the model iteratively generates overlapping clips with a configurable motion frame overlap (default: 9 frames), seamlessly blending segments to produce arbitrarily long video bounded only by the input audio duration and a configurable maximum frame limit.
The diffusion process uses a configurable number of denoising steps (default: 40, tunable from 1–100) with TeaCache acceleration for improved throughput. On NVIDIA H200 hardware, each 81-frame clip requires approximately 3.5 minutes of processing time, yielding a generation-to-output ratio of roughly 10–30× depending on resolution and hardware load.
For 720p output, the system employs a two-stage pipeline: base generation at 480p followed by conditional FlashVSR 4× upscaling (target: 921,600 pixels at 25 fps), applied automatically when audio duration is 60 seconds or less.
4. Performance Highlights
InfiniteTalk addresses a specific niche — audio-driven talking-head video — that differs from general-purpose text-to-video or image-to-video models. Its performance should be evaluated primarily on lip-sync accuracy, identity consistency, and long-form stability rather than visual diversity or cinematic motion range.
| Capability | InfiniteTalk | General I2V Models | Dedicated Lip-Sync Tools |
|---|---|---|---|
| Lip-sync accuracy | Phoneme-level, multi-language | N/A (no audio input) | Word-level, often English-only |
| Maximum duration | Up to 10 minutes (streaming) | 5–15 seconds typical | 30–60 seconds typical |
| Identity preservation | High (audio-anchored per-frame) | Moderate (drift in longer clips) | Moderate |
| Dual-person support | Native | Not available | Rare |
| Resolution | 480p native, 720p with VSR | Up to 1080p | Varies |
| Audio input | Any language WAV/MP3 | N/A | Usually English TTS |
InfiniteTalk achieves strong lip-sync fidelity across Chinese, English, Japanese, and other languages tested, owing to the language-agnostic Wav2Vec2 audio feature extraction. Identity drift is minimal even in 5+ minute generations due to the per-frame audio conditioning anchor.
5. Intended Use & Applications
Digital Avatar & Virtual Presenter: Create realistic talking-head videos for virtual hosts, AI assistants, and digital spokespersons using a single photo and recorded or synthesized speech audio.
Video Dubbing & Localization: Generate lip-synced video from translated audio tracks, enabling cost-effective multilingual content adaptation without re-filming or manual lip-sync editing.
Online Education & Training: Produce instructor-led video content at scale from lecture audio recordings and a single instructor photograph, reducing video production costs for e-learning platforms.
Podcast & Interview Visualization: Transform audio-only podcast or interview recordings into engaging video content with realistic speaker animations, suitable for social media distribution.
Customer Service & Chatbot Video: Generate personalized video responses driven by TTS audio output, enabling human-like video communication in automated customer interaction flows.
Social Media Content at Scale: Rapidly produce talking-head content for influencer accounts, news summaries, or commentary formats using text-to-speech pipelines combined with InfiniteTalk video generation.
