What problem does it solve?
Many production workflows need exact, repeatable control over who speaks, what they say, how scenes are composed, and how outputs integrate into editing pipelines; manually coordinating avatar selection, voice matching, asset uploads, and compositing is slow and error-prone. This Skill centralizes HeyGen video generation best practices so teams can reliably produce brand-consistent avatar videos, transparent WebMs for compositing, and talking-photo presenters at scale.
Core Features & Use Cases
- Precise Avatar & Voice Selection: List and preview HeyGen avatars, use default_voice_id or pick specific voice_ids for guaranteed pairing.
- Multi‑Scene, Multi‑Background Workflows: Build multi-scene videos with different backgrounds per scene, including image and looping video backgrounds.
- Asset Management & Uploads: Upload images, videos, and audio assets and use returned asset URLs or image_keys for backgrounds and talking-photo avatars.
- Transparent WebM & Remotion Integration: Produce WebM outputs with transparency for compositing and follow Remotion best practices (OffthreadVideo, matching dimensions, downloading expiring URLs).
- Talking Photo / Avatar IV Workflows: Create talking-photo avatars from uploads or generate AI photos (Avatar IV) for high-quality single-image avatars.
- Batch, Templates & Webhooks: Support template-driven batch generation, quota checks, polling patterns, and optional webhook notifications for production systems.
Quick Start
Generate a HeyGen avatar video by selecting an avatar_id and voice_id, uploading any custom background assets, invoking the v2/video/generate endpoint with your scene list, and polling for completion while HEYGEN_API_KEY is set in the environment.
