24-ai-avatar-production

What it does

Provides end-to-end guidance for producing AI avatars and AI-generated videos, including single-avatar production, multilingual translation workflows, batch production, and hybrid real+AI formats. Covers reference image intake, face/style/logo/palette replacement, voice cloning, anti-detection tactics for VN platforms, ethics/disclosure, and a QA framework with a 100-point score. It names tools across tiers (Free/Pro/Enterprise) and outlines triggers to start workflows.

How it works

Six-step workflow for a single avatar: 1) Script creation via Vietnamese script template (tied to another skill), 2) Voice generation or clone via ElevenLabs/Vbee/HeyGen Voice, 3) Avatar selection or upload, 4) Render by combining voice and avatar with chosen background and gestures, 5) QA review to meet a 100 diem QA score, 6) Publish export as MP4 to a platform. It provides a 60-second script template with hook/problem/solution/proof/CTA and an output template including Script, Voice, Avatar, Resolution, Duration, QA Score, and Publish details. It describes translating workflows by creating an original Viet video, uploading to a translator tool (e.g., Rask AI), selecting target languages, auto lipsync, and post-review. It details batch production with a 5-day plan: 10 scripts, 10 voices, 10 videos render, QA, and publish; and provides cost estimates per tier. It also covers Hybrid Real+AI by mixing a short real intro with AI body content, plus steps to align visuals (cutting, color matching) in CapCut/Premiere. Finally, it provides a dedicated Voice Clone Protocol with sample length, file format, and consent form; and anti-detection guidance including indicators, fixes, and human-feel enhancements (grain, zoom, color grade, overlays, B-roll, sound design).

When to use it

Trigger phrases include: 'tao avatar AI', 'video AI HeyGen', 'voice clone ElevenLabs', 'lipsync Captions', 'AI video translate', 'talking head AI', 'batch video AI', 'pipeline AI avatar', 'anh reference avatar', 'prompt avatar AI', 'thay mat avatar'. Use workflows when creating a video avatar from zero to publish, translating to multilingual markets, producing multiple videos in a batch, or combining real footage with AI-generated content for efficiency and trust.

What it can touch

Tools referenced for workflows: HeyGen, Synthesia, ElevenLabs, Captions, Rask AI, Vbee. It specifies using these tools across tiers and steps (e.g., voice cloning, avatar rendering, translation, lipsync). It also mentions CapCut and Premiere for hybrid workflows and color grading, plus external QA and publishing pipelines.

Caveats

Mentions ethics and disclosure through VN regulations (Nghi dinh 147/2024). Includes anti-detection practices for VN platforms, which may raise platform-specific risk considerations. QA scoring is mandated (QA Score 100 diem) but no explicit guarantee of outcomes. License is MIT, but no additional usage constraints are provided beyond the skill’s content. The review notes that some workflows rely on external translation and cloning services with associated costs, and that translate lipsync for Vietnamese may be imperfect.