tencentcloud-video-face-fusion_skill

Overview

This skill provides an end-to-end video face fusion capability using Tencent Cloud's video creation APIs. It extracts core facial features from a user photo and a template video, then generates a natural-looking fused video with optional branding. The package includes scripts for one-step generation, job submission, and job query to support asynchronous workflows.

How this skill works

The skill calls Tencent Cloud Video Face Fusion APIs to submit a fusion job, poll for completion, and retrieve the resulting video. It accepts a template video URL and one or more template face images plus corresponding user face images, maps faces one-to-one, and applies style-adaptive blending. Scripts handle authentication via environment variables, dependency checks, and optional parameters like logo watermarking and polling controls.

When to use it

• Create promotional or branded videos where a user’s face is swapped into a template scene.
• Automate bulk or multi-face fusion jobs in marketing campaigns or social features.
• Prototype entertainment UGC features, reels, or interactive social content with high shareability.
• Integrate into an asynchronous pipeline that needs job submission and later status polling or retrieval.

Best practices

• Set Tencent Cloud SecretId and SecretKey as environment variables; never hard-code keys in code.
• Use the main.py one-step script for most cases to submit and auto-poll until completion.
• Match the number and order of --template and --face inputs for correct one-to-one replacement.
• Keep template video constraints in mind (≤1GB, ≤20s, ≤4K, fps≤25) to avoid API rejections.
• Enable max-poll-time and poll-interval to balance responsiveness and API rate limits; handle API errors gracefully and surface clear failure messages to users.

Example use cases

• Single-face marketing clip: swap a customer’s face into a short product demo video for personalized ads.
• Multi-face group scene: replace multiple template faces with different user photos for team or family videos.
• Asynchronous batch processing: submit many jobs with submit_job.py and query results later in a scheduler.
• Interactive social app: integrate main.py into a server endpoint to produce shareable fused videos on demand.