Overview AI Talking Photo brings static photos to life by animating faces to speak with realistic lip-sync and natural facial movements. The API analyzes facial features and synchronizes mouth movements, head poses, and expressions with provided audio or generated speech.
How It Works
Provide a photo - Upload an image with a clear faceAdd audio - Upload audio or provide text for speech generationAPI animates - AI creates realistic lip-sync and facial movementsDownload video - Retrieve your animated talking photo
Use Cases
Marketing videos - Create spokesperson videos from headshotsEducational content - Animate historical figures or charactersPersonalized messages - Send video messages from static photosSocial media - Create engaging content from profile picturesPresentations - Add dynamic talking heads to slides
Best Practices Photo Selection Use clear, front-facing photos - Best results come from high-quality headshots with visible
facial features.
Good lighting - Well-lit faces produce better animationsFront-facing angles - Avoid extreme profile shotsClear features - Eyes, nose, and mouth should be unobstructedHigh resolution - At least 512x512 pixels recommended
Audio Guidelines Audio Type Best Practice Voice recording Clear speech without background noise Generated speech Use natural-sounding text prompts Music/songs Works best with clear vocals Length Keep under 30 seconds for best results
Code Examples Basic Talking Photo with Text from magic_hour import Client from os import getenv client = Client( token = getenv( "API_TOKEN" )) result = client.v1.ai_talking_photo.generate( assets = { "image_file_path" : "https://raw.githubusercontent.com/runshouse/Sample_Assets/main/tomcruise.png" , "audio_file_path" : "https://raw.githubusercontent.com/runshouse/Sample_Assets/main/you-are-just-a-line-of-code.mp3" }, name = "Talking Photo" , start_seconds = 0 , end_seconds = 2 , wait_for_completion = True , download_outputs = True , download_directory = "outputs" ) if result.status == "complete" : print ( f "✅ Talking photo complete!" ) print ( f "Downloaded to: { result.downloaded_paths } " ) print ( f "Credits charged: { result.credits_charged } " ) else : print ( f "❌ Job failed with status: { result.status } " ) if hasattr (result, 'error_message' ): print ( f "Error: { result.error_message } " )
With Audio File result = client.v1.ai_talking_photo.generate( assets = { "image_file_path" : "https://raw.githubusercontent.com/runshouse/Sample_Assets/main/tomcruise.png" , "audio_file_path" : "https://raw.githubusercontent.com/runshouse/Sample_Assets/main/you-are-just-a-line-of-code.mp3" }, name = "Talking Photo" , wait_for_completion = True , download_outputs = True , download_directory = "outputs" ) if result.status == "complete" : print ( f "✅ Talking photo complete!" ) print ( f "Downloaded to: { result.downloaded_paths } " ) print ( f "Credits charged: { result.credits_charged } " ) else : print ( f "❌ Job failed with status: { result.status } " ) if hasattr (result, 'error_message' ): print ( f "Error: { result.error_message } " )
Pricing Talking Photo pricing varies by video length and resolution:
Configuration Credits per Second 720p or lower ~10-15 credits/sec Higher resolution ~20-30 credits/sec
Resolution Limits AI Talking Photo has a maximum resolution of 720p across all subscription tiers due to computational requirements.
API Reference
AI Talking Photo API Reference View full API specification
