Skip to main content
Talking Portrait Node

​
Talking Portrait

Make any portrait speak.
This node takes a still image of a face and syncs it to audio, creating natural lip movement and subtle facial animation. Perfect for bringing characters to life or creating talking avatars.

​
🪄 When to Use It

Use Talking Portrait when you want to animate a still face with speech. It works great for:
  • Character dialogue — make your illustrated characters speak
  • Avatar creation — animated profile pictures or mascots
  • Story narration — portraits that deliver voiceover
  • Explainer videos — talking head style without filming
Skill level: Intermediate
Average time: 30–60 seconds per clip
Cost: Medium-High

​
🎚️ Controls and Parameters

​
Image (Image, Required)

A portrait or character image with a visible face. Works best with:
  • Clear frontal or 3/4 view of the face
  • Good lighting on the face
  • Minimal motion blur
💡 Tip: AI-generated portraits often work better than photos because they’re cleaner and more consistent.

​
Audio (Audio, Required)

The speech or dialogue you want the portrait to lip-sync to. Can be:
  • Generated from Voice Generator node
  • Uploaded audio file
  • Connected from another audio node

​
Prompt (Text, Required)

Describe the character and context — helps the AI understand the mood and style. 💡 Tip: “Professional woman speaking confidently to camera” or “Animated character telling a story” sets the tone.

​
Seed (Number)

Control randomness. Same seed + same inputs = same result.

​
Intensity (Slider, 0–1)

How much facial movement to add:
  • Low (0.3–0.5) — subtle, natural movement
  • Medium (0.6–0.8) — noticeable but realistic
  • High (0.9–1.0) — expressive, animated style
💡 Tip: Start around 0.8 for most portraits. Lower for more realistic, higher for cartoon-like.

​
🎨 Available Models

​
Infinite Talk (Default)

Advanced lip-sync technology optimized for portrait animation. Produces natural facial movement with accurate lip-sync to audio. Supports:
  • Variable intensity control
  • Seed-based consistency
  • Custom prompts for context
💡 Tip: Infinite Talk is currently the only model for this node and handles all portrait animation scenarios.

​
🎨 What to Expect

The AI will:
  • Animate the mouth to match the audio
  • Add subtle head movements and eye blinks
  • Keep the overall character appearance consistent
Works best with:
  • Clear audio — crisp dialogue without background noise
  • Forward-facing portraits — frontal or slight angle views
  • Well-lit images — shadows on the face can cause issues
Less predictable with:
  • Profile views or extreme angles
  • Multiple faces in one image (picks the main face)
  • Very stylized or abstract art

​
💬 Quick Tips

  • Generate portraits with Text to Image for most consistent results
  • Use Voice Generator to create matching audio clips
  • Keep intensity around 0.7-0.9 for natural-looking motion
  • Audio length determines video length (longer audio = longer processing)
  • Works with illustrated characters, not just realistic portraits