Dream Avatar

DreamAvatar - Digital Human Video Generator

Generate talking avatar videos from images and audio using DreamAvatar 3.0 Fast API.

Quick Start

1. Get API Key

Before using this skill, you need a DreamAPI API key. Visit https://api.newportai.com/api-reference/get-started to sign up and get your API key.

For more AI tools, please visit: https://tools.dreamfaceapp.com/home

2. Configure API Key

Users must set their DreamAPI key in OpenClaw config:

openclaw config patch --json '{"skills": {"entries": {"dream-avatar": {"env": {"DREAM_API_KEY": "your-api-key-here"}}}}}'

Or add to ~/.openclaw/openclaw.json:

{
  "skills": {
    "entries": {
      "dream-avatar": {
        "env": {
          "DREAM_API_KEY": "your-api-key"
        }
      }
    }
  }
}

3. Generate Video

You can provide input in two ways:

Option A: Local Files (Recommended)

• image: Path to a local image file (jpg, jpeg, png, webp, gif)
• audio: Path to a local audio file (mp3, wav, mp4, max 3 minutes)
• prompt: Description of the expression/behavior
• resolution (optional): "480p" or "720p", default "480p"

The skill will automatically upload your local files and generate the video.

Option B: Public URLs

• image: URL to a publicly accessible image (jpg, jpeg, png, webp, gif)
• audio: URL to a publicly accessible audio (mp3, wav, mp4, max 3 minutes)
• prompt: Description of the expression/behavior
• resolution (optional): "480p" or "720p", default "480p"

API Details

1. Get Upload Policy

POST https://api.newportai.com/api/file/v1/get_policy

Headers:

Authorization: Bearer {DREAM_API_KEY}
Content-Type: application/json

Request Body:

{
  "scene": "Dream-CN"
}

Response:

{
  "code": 0,
  "message": "success",
  "data": {
    "accessId": "LTAI5tF1QzxoHGvEcziVACyc",
    "policy": "eyJ0...",
    "signature": "G2TzrhlybemHbfFakysY4j2EI2I=",
    "dir": "tmp/dream/2024-11-19/5369207820989002/",
    "host": "https://...",
    "expire": "1732005888",
    "callback": "eyJ0..."
  }
}

2. Upload File to OSS

Form Data:

Response:

{
  "code": 0,
  "message": "success",
  "data": {
    "verifyStatus": false,
    "reqId": "732c9caa-0a2e-4aa1-87d9-52430a8f0314"
  }
}

The uploaded file URL will be: {host}/{key}

3. Generate Video (Image to Video)

POST https://api.newportai.com/api/async/dreamavatar/image_to_video/3.0fast

Headers:

Authorization: Bearer {DREAM_API_KEY}
Content-Type: application/json

Request Body:

{
  "image": "https://.../photo.jpg",
  "audio": "https://.../speech.mp3",
  "prompt": "a person smiling and speaking",
  "resolution": "480p"
}

Response (Task Submission):

{
  "code": 0,
  "message": "success",
  "data": {
    "taskId": "aa4bf173ffd84c2f8734d536d6a7b5a7"
  }
}

4. Polling for Result

Use the taskId to poll for results:

POST https://api.newportai.com/api/getAsyncResult

Request body:

{
  "taskId": "your-task-id"
}

Response When Complete:

{
  "code": 0,
  "message": "success",
  "data": {
    "task": {
      "taskId": "aa4bf173ffd84c2f8734d536d6a7b5a7",
      "status": 3,
      "taskType": "dreamavatar/image_to_video/3.0fast"
    },
    "videos": [
      {
        "videoType": "mp4",
        "videoUrl": "https://...video.mp4"
      }
    ]
  }
}

Implementation Notes

• Local files: The skill handles upload automatically. Files are uploaded to OSS and URLs are generated.
• URL input: Image and audio URLs must be publicly accessible (not behind authentication)
• Audio duration: Cannot exceed 3 minutes
• The API is async: You must poll for results
• Poll every 2-3 seconds, with a timeout of ~60 seconds
• Status codes: 0 = pending, 1 = processing, 2 = processing (not failed!), 3 = completed, 4 = timeout
• Uploaded file URLs: Only valid for 1 day (Storage service is free but for AI program support only)

Upload Process Flow

Local File → Get Upload Policy → Upload to OSS → Get Public URL → Generate Video

The skill implements this flow automatically when local files are provided:

1. Call get_policy to receive OSS credentials
2. Upload image/audio to OSS
3. Construct public URL from host + key
4. Call DreamAvatar API with the uploaded URLs
5. Poll for video completion
6. Return the final video URL