> ## Documentation Index
> Fetch the complete documentation index at: https://docs.wrldwide.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Live Avatar
> Generate talking avatar videos from portrait photos and audio — 6 models including OmniHuman v1.5 and Kling Avatar v2
# Live Avatar
Turn any portrait photo into a talking, animated avatar. Upload a photo and an audio file — the model generates a synchronized talking head video with natural facial expressions and head movements.
**Credit Cost:** 8–20 credits (model-dependent)\
**Access:** [app.wrldwide.ai/live-avatar](https://app.wrldwide.ai/live-avatar)
***
## Available Models
### OmniHuman v1.5 — 15 credits ⭐ BEST
**Top-quality avatar generation with superior realism**
* **Input:** Portrait photo + audio file
* **Output:** Talking avatar video
* **Character:** Most natural expressions, realistic head movements, best lip sync accuracy
* **Best for:** Professional brand avatars, spokesperson videos, polished content
***
### Kling Avatar v2 Pro — 12 credits
**Kling's premium avatar model**
* **Input:** Portrait photo + audio file
* **Output:** High-quality talking avatar
* **Character:** Professional quality, good expression range
* **Best for:** Brand content where OmniHuman budget is tight
***
### Kling Avatar v2 — 8 credits
**Kling's standard avatar model**
* **Input:** Portrait photo + audio file
* **Output:** Talking avatar video
* **Character:** Good quality, reliable performance
* **Best for:** Testing, iteration, batch production
***
### Live Portrait (Video) — 12 credits
**Drive avatar with a reference video's expressions**
* **Input:** Portrait photo + driver video (video of a real person expressing)
* **Output:** Portrait animated with the expressions and movements from the driver video
* **Character:** Transfers exact expression timing from driver video
* **Best for:** When you want specific expression choreography, not just audio sync
***
### Live Portrait (Audio) — 15 credits
**Live Portrait driven by audio**
* **Input:** Portrait photo + audio file
* **Output:** Portrait animated to audio
* **Character:** Natural motion, high quality
* **Best for:** High-quality audio-driven avatar generation
***
### SadTalker — 20 credits
**Advanced avatar with expression scale control**
* **Input:** Portrait photo + audio file
* **Output:** Talking avatar with controllable expression intensity
* **Extra setting:** Expression Scale (slider) — controls how expressive the avatar becomes
* **Character:** Detailed expression control, distinct style
* **Best for:** When you need fine-grained control over expression intensity
***
## How to Use
Navigate to [app.wrldwide.ai/live-avatar](https://app.wrldwide.ai/live-avatar)
Choose based on your quality needs and mode:
* **Best quality** → OmniHuman v1.5
* **Balanced quality/cost** → Kling Avatar v2 Pro
* **Budget/iteration** → Kling Avatar v2
* **Expression choreography** → Live Portrait (Video)
* **Expression control** → SadTalker
Upload a clear portrait photo of the avatar subject.
**Best results:**
* Frontal or near-frontal face angle
* Clear, well-lit face
* Neutral expression (model will add the expressions)
* Single person, no occlusions over the face
* **Audio-driven models** → Upload MP3, WAV, or M4A
* **Live Portrait (Video)** → Upload an MP4 video of a real person expressing
If using SadTalker, set the Expression Scale slider (1.0 = default, higher = more expressive)
Click **Generate**. Processing takes 1–4 minutes depending on audio length and model.
Watch the generated avatar video. Download for use in ads, social media, or brand content.
***
## Model Comparison
| Model | Cost | Input | Quality | Best For |
| ------------------------- | ---- | ------------- | ------- | ----------------------- |
| **OmniHuman v1.5** | 15cr | Photo + Audio | ⭐⭐⭐⭐⭐ | Professional/brand |
| **Kling Avatar v2 Pro** | 12cr | Photo + Audio | ⭐⭐⭐⭐ | Premium quality |
| **Kling Avatar v2** | 8cr | Photo + Audio | ⭐⭐⭐ | Testing/iteration |
| **Live Portrait (Video)** | 12cr | Photo + Video | ⭐⭐⭐⭐ | Expression choreography |
| **Live Portrait (Audio)** | 15cr | Photo + Audio | ⭐⭐⭐⭐ | High-quality audio sync |
| **SadTalker** | 20cr | Photo + Audio | ⭐⭐⭐⭐ | Expression control |
***
## Use Cases
### Brand Spokesperson
Create a branded AI spokesperson — upload a generated or licensed portrait, add a professional voice-over, and produce a consistent spokesperson for ads, onboarding, or product demos.
### Multilingual Content
Record voice-overs in 5 languages, generate the same avatar speaking each language. Localize video content at scale without hiring multi-language talent.
### UGC-Style Talking Heads
Generate UGC-style talking head videos for ads without filming. Combine with Face Swap to customize the avatar appearance.
### Product Demo Narration
Place a talking avatar next to a product animation to narrate features. Combine with Product Animation for a complete product showcase video.
***
## Tips for Best Results
**Portrait Photo:**
* Neutral background is ideal (can remove background first with [Background Removal](/en/features/background-removal))
* Eyes open, looking roughly toward camera
* Good natural lighting, avoid harsh flash
* At least 512×512 resolution
**Audio:**
* Studio or near-studio quality is ideal
* Clear enunciation
* Minimal background noise
* Keep pauses natural (the avatar will pause movements to match)
***
## Next Steps
Sync audio to existing video content
Swap faces in videos
Full UGC ad production workflow
Complete UGC video production wizard