B
77.0Overall score
A setup that takes a written script, generates natural voiceover, and drives a lip-synced avatar so you can publish faceless explainer videos on a schedule. For solo educators and faceless channels that want volume without filming.
77.0Score
880Votes
5Components
Install this build
terminal
python avatar_pipeline.py --script topic.txtComponents
Model
- Claude Haiku 4.6 for scripts
- XTTS v2 for voice
Stack
- SadTalker for lip-sync
- FFmpeg
- Python orchestrator
Hardware
- 8GB VRAM GPU
- CPU works but renders slowly
How it works
- Expand a topic into a tight spoken script
- Synthesize voiceover with a cloned or stock voice
- Drive the avatar image with the audio track
- Mux audio, video and B-roll into the final cut
Rules
- Keep scripts conversational, short sentences only
- Always disclose the video is AI generated
Summary
A setup that takes a written script, generates natural voiceover, and drives a lip-synced avatar so you can publish faceless explainer videos on a schedule. For solo educators and faceless channels that want volume without filming.
77.0 score 880 votes
0 Reviews
Your rating
Sign in to post
Loading discussion...