Aurora Avatar

Fresh

Purpose

Generate studio-grade, hyper-realistic avatar videos from a single image using Creatify's state-of-the-art Aurora model.

Prerequisites

Authenticated API access (Authentication SOP)
A single high-quality portrait image (or image URL)
Text script or audio URL
Sufficient credits

What Makes Aurora Different

Aurora is Creatify's SOTA (state-of-the-art) avatar model. Unlike v1/v2 lipsync which uses pre-made avatar templates, Aurora takes a single image and produces production-quality, lifelike avatar video with:

Natural head movements and micro-expressions
Studio-grade lighting and rendering
Minimal artifacts compared to standard lipsync

Procedure

Step 1: Prepare Your Image

Requirements for the input image:

Clear, front-facing portrait
Good lighting, minimal shadows
High resolution (minimum 512x512 recommended)
Single person in frame
Publicly accessible URL

Step 2: Create an Aurora Task

bash

curl --request POST \
  --url https://api.creatify.ai/api/aurora/ \
  --header "Content-Type: application/json" \
  --header "X-API-ID: $CREATIFY_API_ID" \
  --header "X-API-KEY: $CREATIFY_API_KEY" \
  --data '{
    "image_url": "https://example.com/portrait.jpg",
    "text": "Welcome to our company. I am excited to share our latest innovations with you today.",
    "voice_id": "voice-uuid-001",
    "aspect_ratio": "9:16"
  }'

Key parameters:

Parameter	Type	Required	Description
`image_url`	string	Yes	URL of the portrait image
`text`	string	Yes (or `audio_url`)	Script for the avatar to speak
`audio_url`	string	Yes (or `text`)	Pre-recorded audio URL
`voice_id`	string	Yes (with `text`)	Voice for text-to-speech
`aspect_ratio`	string	No	`9:16`, `16:9`, or `1:1`

Response:

json

{
  "id": "aurora-task-uuid",
  "status": "pending",
  "image_url": "https://example.com/portrait.jpg"
}

Step 3: Poll for Completion

bash

curl --request GET \
  --url https://api.creatify.ai/api/aurora/aurora-task-uuid/ \
  --header "X-API-ID: $CREATIFY_API_ID" \
  --header "X-API-KEY: $CREATIFY_API_KEY"

TIP

Aurora processing may take longer than standard lipsync due to the higher quality rendering. Poll every 10-15 seconds.

Completed response:

json

{
  "id": "aurora-task-uuid",
  "status": "done",
  "output": "https://cdn.creatify.ai/videos/aurora-output.mp4",
  "duration": 20
}

Step 4: Download Output

Download the video from the output URL.

Credit Cost

Operation	Credits
List aurora items	0
Create Aurora task	Varies (SOTA pricing)

WARNING

Aurora uses the SOTA model which may have different credit costs than standard lipsync. Check the Credit Costs Reference for current pricing.

Aurora vs Lipsync v1 vs v2

Feature	Lipsync v1	Lipsync v2	Aurora
Input	Avatar ID	Avatar IDs	Any image
Quality	Good	Good	Studio-grade
Scenes	1	Multiple	1
Realism	Template-based	Template-based	Generative
Processing time	Fast	Medium	Longer
Use case	Quick avatar videos	Multi-scene narratives	Premium content

Verification Checklist

[ ] Input image is high quality and front-facing
[ ] Image URL is publicly accessible
[ ] Task creation returns pending status
[ ] Final status is done with valid output URL
[ ] Video shows natural head movements and expressions
[ ] Lip sync matches the text/audio accurately
[ ] No visual artifacts around face/hair

Common Errors

Error	Cause	Fix
Image not accessible	Private or broken URL	Use a publicly accessible image URL
Low quality output	Input image too small or blurry	Use a higher resolution, well-lit portrait
Face not detected	Image does not contain a clear face	Use a front-facing portrait photo

Aurora Avatar ​

Purpose ​

Prerequisites ​

What Makes Aurora Different ​

Procedure ​

Step 1: Prepare Your Image ​

Step 2: Create an Aurora Task ​

Step 3: Poll for Completion ​

Step 4: Download Output ​

Credit Cost ​

Aurora vs Lipsync v1 vs v2 ​

Verification Checklist ​

Common Errors ​