HeyGen article examples (50590)
HeyGen article examples (50590)
End-to-end workflow for a HeyGen talking-head video, Whisper transcript timing, slide deck layouts, PiP calibration, and five media variant MP4 exports.
Overview of all new video/PiP features: Recent features.
Full pipeline (video + transcript → YAML → PPTX → MP4), automation gaps, and checklists: Video + transcript workflow.
Repository paths are under examples/ in the PraisonAIPPT repo.
Workflow overview
flowchart LR
A[heygen-50590-content.yaml] --> B[sync_heygen_variants.py]
B --> C[Variant YAMLs]
C --> D[praisonaippt build]
D --> E[PPTX + slide JPEGs]
D --> F[MP4 export]
| Step | Action |
|---|---|
| 1 | Edit examples/heygen-50590-content.yaml — slide copy, slide_type, duration_sec, notes, avatar_video_path |
| 2 | Run python examples/sync_heygen_variants.py — copies content into five variant YAMLs with correct narration_mode / audio_path |
| 3 | Build PPTX and/or MP4 per variant (below) |
Presenter narration for captions lives in each verse notes field. Slide headlines and deck layouts follow the Claude Managed Agents article (May 2026).
Shared assets (~57 s)
| File | Role |
|---|---|
examples/heygen-article-50590.mp4 |
HeyGen headshot + embedded AAC voice |
examples/short-script-50590.mp3 |
Separate narration MP3 (Whisper / trim source) |
examples/short-script-50590_timestamps.json |
Whisper transcript + segment timing |
examples/short-script-50590_timestamps.txt |
Word-level timestamps (reference) |
examples/heygen-article-50590-words.srt |
Karaoke-style word captions |
Narration source (pick one)
- Default — HeyGen video audio —
narration_mode: avataroraudio_source: heygen_video. Lip-sync / TTS from the MP4 track. - Optional — video + separate MP3 —
narration_mode: audio_fileoraudio_source: external. PiP video is visual-only for audio; MP3 drives timing and voice. - TTS —
narration_mode: tts. Synthesise from versenotes; avatar video is muted for narration.
With narration_mode: auto, HeyGen embedded audio wins when the avatar file has a track, even if audio_path is also set on the verse. See Video export — narration modes.
Variant matrix
| YAML | Video (HeyGen PiP) | Audio source | narration_mode |
Typical output |
|---|---|---|---|---|
heygen-50590-video-audio-heygen.yaml |
Yes | HeyGen MP4 | avatar |
Default talking-head export |
heygen-50590-video-visual-mp3.yaml |
Yes (muted for audio) | MP3 | audio_file |
Studio / re-recorded voiceover |
heygen-50590-audio-only.yaml |
No | MP3 | audio_file |
Slides + voiceover only |
heygen-50590-video-only-silent.yaml |
Yes | None | fixed |
B-roll / timing preview |
heygen-50590-slides-silent.yaml |
No | None | fixed |
Slide timing demo |
Images variant (full-bleed product heroes)
Separate from the five media variants — same HeyGen MP4 + Whisper timings as heygen-50590-video-audio-heygen.yaml, with product screenshots and floating headline panels:
| File | Role |
|---|---|
heygen-50590-video-audio-heygen-images.yaml |
Full-bleed avatar_media_3, per-slide text_panel.anchor |
heygen-50590-video-audio-heygen-images.pptx / .mp4 / .srt |
Built outputs (~57 s) |
slide_images/heygen-50590-images/ |
Layout JPEGs + golden/ + mp4-frames/ |
# Key settings (see file for full deck)
skip_title_slide: true
slide_style:
layouts:
avatar_media_3:
hero_layout: full_bleed
text_style: semi_panel
pipeline:
golden_slide_dir: slide_images/heygen-50590-images/golden
export_mp4_frames: true
praisonaippt -i examples/heygen-50590-video-audio-heygen-images.yaml \
-o examples/heygen-50590-video-audio-heygen-images.pptx \
--convert-video --video-output examples/heygen-50590-video-audio-heygen-images.mp4
praisonaippt validate-deck -i examples/heygen-50590-video-audio-heygen-images.yaml
Docs: Avatar layouts — full-bleed hero · Slide QA
Build one variant
VARIANT=heygen-50590-video-audio-heygen
python -m praisonaippt.cli \
-i examples/${VARIANT}.yaml \
-o examples/${VARIANT}.pptx \
--convert-video \
--video-output examples/${VARIANT}.mp4 \
--no-list-slides
video_export.narration_mode in each variant YAML is applied automatically; no extra CLI flag is required.
Build all showcase outputs
python examples/sync_heygen_variants.py
python examples/build_showcase_examples.py
Rebuilds avatar gallery, deck gallery, and all five HeyGen variant PPTX + MP4 files.
Generate variant YAMLs from Whisper JSON
python -m praisonaippt.cli transcript-to-yaml \
-i examples/short-script-50590_timestamps.json \
-o examples/heygen-article-50590 \
--variants all
Writes combination decks (same presets as MEDIA_VARIANTS in code). For the maintained article deck, prefer content + sync above.
Related configuration
| Topic | Doc |
|---|---|
PiP face centre / crop_x |
Avatar PiP calibration |
avatar_quote (video-only PiP, no double avatar) |
Avatar layouts |
| Full-bleed hero screenshots | Avatar layouts — avatar_media_3 |
| Slide JPEG + golden + MP4 frame QA | Slide QA · Slide JPEG export |
video_export keys |
YAML deck reference |
| CLI flags | Commands — video & avatar |
Legacy filenames
| File | Equivalent variant |
|---|---|
heygen-article-50590-short.yaml |
video-audio-heygen (hand-tuned headlines) |
heygen-article-50590-short-audio-only.yaml |
audio-only |
Source of truth in repo
The canonical copy for editors is examples/heygen-50590-examples.md (mirrors this page). After editing content, run sync_heygen_variants.py before building variants.