YouTube Publish (Scripted Flow)

Use scripts in order. Stop for validation after copy + thumbnail generation. Ask exact publish time if not provided.

Behavior rules for the agent

• Tone & Authority: Titles and copy must focus on engineering, architecture, and solving developer friction. Avoid reaction-style hype and mass-content phrasing.
• Title Blacklist (strict): Forbidden in titles: RIP, Increíble, Brutal, Locura, Definitivo, ¿El fin de...?, and crown/fire emojis.
• Technical Anchor (strict): Every title must include at least one engineering keyword: Orquestación, Despliegue, Infraestructura, Clean Architecture, Refactorización, Pipeline, Capa de Abstracción.
• Title Derivation: Do not ask for a title hint; derive it from the video stem and the technical density of the SRT.
• Scheduling: If the user provides a publish time, resolve to exact YYYY-MM-DD HH:MM using system time and pass --publish-at + --timezone. Always determine and pass --timezone.
• Content Generation Engine: Titles, thumbnail ideas, description, chapters, and LinkedIn copy must be written by the same model executing this skill.
• English Variant: When the user wants a multilingual YouTube version, always generate an English pack for the same video: translated transcript, English title, English description, and dubbed English audio.
• English Scope: The English pack is for YouTube multi-language audio/localization only. Do not create English social posts unless the user explicitly asks for them.
• Thumbnail Generation: Generate 3 thumbnails using the presenter photo set. Default presenter is antonio (assets/antonio-1.png, antonio-2.png, antonio-3.png). If the user indicates the video is from Nino, switch presenter to nino (assets/nino-1.png, nino-2.png, nino-3.png). Keep only two non-negotiables: (1) massive bold white text (max 3-4 words), (2) cinematic dark look with cyan/magenta accents. Everything else should adapt to the video's narrative with maximum creative freedom.
• Reference Photos (strict): For each generated thumbnail, always pass the 3 presenter images together as references. They are identity anchors (not fixed poses); the model is free to choose the best posture/composition.
• Thumbnail Engine: Reuse 3rd-nano-banana-pro/scripts/generate_image.py as the single image generation engine. Keep default model behavior (Flash). Only override model explicitly when requested.
• Thumbnail Creativity Rule: Deliver 1 safer option + 2 exploratory options. Avoid producing near-duplicates. If an unconventional concept communicates better for that specific video, prioritize it.
• Workflow: Upload a private draft before generating copy so the video URL can be used in social text.
• Newsletter: Disabled in this flow. Do not generate or schedule newsletter here.
• X Strategy: Do not schedule/publish to X via PostFlow in this flow. X is handled as native video upload outside this step.
• Links: In social posts, the comment must not be just the link; it must include a brief descriptive text inviting to watch (e.g., "Watch the full technical analysis here: https://...").
• Comment Sequence: For final publish/update, always use this order: set video to unlisted, insert promo comment (Domina la IA...), then set final status (private with publishAt if scheduled, otherwise private).
• Schedule Decision Required: Never publish without an explicit decision in Programación (final): either a date YYYY-MM-DD HH:MM or private.
• Timing: Schedule social posts 15 minutes after the YouTube publish time.
• Audio Consistency: prepare_video.py runs audio normalization in auto mode by default. It analyzes LUFS/true peak/LRA and only re-encodes audio when out of target.

Content Styles

LinkedIn Post Style

• Length/Format: 600–900 characters, 3–6 short paragraphs, 1–2 emojis.
• Strategy (Signal vs Noise): Start with a principle or real engineering problem, not with "new video" framing.
• Identity: Keep the tone of technical authority. Fewer creator-marketing phrases, more architecture conclusions and tradeoffs.
• Scope: 1 central idea focused on technical authority. No digressions.
• Closing: Final line “Link en el primer comentario.” followed by a short question or CTA.
• Restrictions: No hashtags.

Scripted flow (order)

1. Prepare video

   • Command:
     bashCopy

     python scripts/prepare_video.py --videos /path/v1.mp4 [/path/v2.mp4 ...]
     

   • Audio behavior (default): --audio-normalization auto targets -14 LUFS, -1 dBTP, and max LRA 9; it skips normalization when already in range.
   • Optional:
     • --audio-normalization always to force normalization.
     • --audio-normalization off to skip analysis and normalization.
   • Output JSON with workdir, video, slug.

2. Upload draft (private)

   • Command:
     bashCopy

     python scripts/upload_draft.py --video <video> --output-video-id <workdir>/video_id.txt --client-secret <path>
     

   • Write video_id.txt and create video_url.txt.

3. Transcribe + clean

   • Command:
     bashCopy

     python scripts/transcribe_parakeet.py --video <video> --out-dir <workdir>
     

   • Outputs transcript.es.cleaned.srt.
   • After transcription, copy the SRT to the vault transcripts folder:
     bashCopy

     cp <workdir>/transcript.es.cleaned.srt ~/Documents/aipal/transcripts/<YYYY-MM-DD>-<slug>.srt
     

   • Use today's date and the video slug from step 1 as the filename.

4. Prepare English dubbing assets (when multilingual output is requested)

   • Read <workdir>/transcript.es.cleaned.srt and create:
     • <workdir>/transcript.en.srt translated to natural English while preserving timestamps.
     • <workdir>/title.en.txt with 1 English YouTube title for the dubbed track.
     • <workdir>/description.en.txt with 1 English YouTube description for the dubbed track.
   • Generate dubbed English audio using the youtube-dubber project.
   • Default dubbing path:
     • scripts/dub_voxtral.py
     • model voxtral-mini-tts-latest
     • English reference clip from the presenter's own voice when available
   • Fallback path:
     • Chatterbox / Qwen only if Voxtral is unavailable or clearly worse for a specific run
   • Save at least:
     • <workdir>/dubbed_audio.en.wav
     • <workdir>/dubbed_video.en.mp4 if the dubbing pipeline also muxes the video
   • Keep the English title/description technically faithful to the Spanish source, not marketing-localized beyond what is needed for natural English.
   • The goal is to run Voxtral through the same timed dubbing pipeline as the other models, not a manual narration-only shortcut.

5. Generate copy with the calling model

   • Read <workdir>/transcript.es.cleaned.srt directly and generate:
     • 3 Technical Authority Titles.
     • 3 Thumbnail ideas (Artifact-based).
     • Description (remove any self-link to current video).
     • Chapters (MM:SS).
     • LinkedIn post (per rules).
   • Save the result into <workdir>/content.md.
   • Also save thumbnail concepts into <workdir>/ideas.json with this shape:
     jsonCopy

     {
       "titles": ["...", "...", "..."],
       "thumbnails": [
         {"text": "...", "artifact": "...", "concept": "..."},
         {"text": "...", "artifact": "...", "concept": "..."},
         {"text": "...", "artifact": "...", "concept": "..."}
       ]
     }
     

   • Make sure content.md contains at least these sections so downstream validation stays compatible:
     mdCopy

     # Pack YouTube — <slug>
     
     ## Enlace del vídeo
     <video_url>
     
     ## Títulos
     - ...
     - ...
     - ...
     
     ## Ideas de thumbnails
     1. Texto: ...
        Artifact: ...
        Concept: ...
     
     ## Descripción
     ...
     
     ## Capítulos
     00:00 ...
     
     ## LinkedIn
     ...
     
     ## Título (final)
     
     ## Descripción (final)
     
     ## Capítulos (final)
     
     ## Post LinkedIn (final)
     
     ## Thumbnail (final)
     
     ## Programación (final)
     (YYYY-MM-DD HH:MM o "private")
     
     ## Title (EN)
     
     ## Description (EN)
     

   • Title quality gate: reject title candidates that break blacklist or technical-anchor rules.

6. Generate 3 thumbnails

   • Use presenter photos according to context: by default antonio, and nino only when explicitly requested for a Nino video. Create 3 images into <workdir>/thumb-1.png, thumb-2.png, thumb-3.png.
   • The same model executing the skill must derive the 3 image prompts from the transcript and ideas.json, then save each prompt into <workdir>/thumb-1.prompt.txt, thumb-2.prompt.txt, thumb-3.prompt.txt.
   • Keep the two anchors fixed (massive white text + cinematic cyan/magenta look), but allow concept/composition/artifact/background to vary freely by story.
   • Target mix: 1 safe option + 2 exploratory options.
   • Example with multiple inputs:
     bashCopy

     uv run /path/to/nano-banana-pro/scripts/generate_image.py --prompt "Antonio working..." --filename "thumb-1.png" --input-image assets/antonio-1.png assets/antonio-2.png assets/antonio-3.png
     

   • If using helper scripts:
     • Batch render from an existing ideas.json: python scripts/generate_missing_thumbs.py --presenter antonio --out-dir <videos-root>
     • Nino video: python scripts/generate_missing_thumbs.py --presenter nino --out-dir <videos-root>
     • Optional override: --image-model <model> only if explicitly needed; otherwise keep the default model.

7. Stop to ask for validation of:

   • Title (choose one of the 3 generated).
   • Thumbnail (choose one of the 3 generated).
   • Description (edit if needed).
   • Chapters (edit if needed).
   • LinkedIn post (edit if needed).
   • English title (edit if needed).
   • English description (edit if needed).

8. Update YouTube

   • Command:
     bashCopy

     python scripts/update_youtube.py --video-id <id> --title "..." --description-file <desc.txt> --thumbnail <thumb.png> --publish-at "YYYY-MM-DD HH:MM" --timezone <IANA> --client-secret <path>
     

9. Build native X video variant (after thumbnail choice)

   • Command:
     bashCopy

     python scripts/build_x_native_video.py --video <video.mp4> --thumbnail <thumb.png> --output <workdir>/video-x.mp4 --intro-ms 500
     

   • Result: a version ready for X where the first 500ms shows the selected thumbnail as a static cover frame.

10. Schedule socials (PostFlow, excluding X)

• Command:
  bashCopy

  python scripts/schedule_socials.py --text-file <linkedin.txt> --scheduled-date <ISO8601+offset> --comment-url <video_url> --image <thumb.png>
  

• This script publishes to configured socials except X.
• Note: schedule_socials.py percent-encodes underscores in the --comment-url (e.g. _ -> %5F) to avoid LinkedIn URL formatting issues.

11. Final Reminder

• Explicitly remind the user to go to YouTube Studio to:
  • Enable monetization (not supported via API).
  • Add End Screens (not supported via API).
  • If multilingual output was requested, upload the English dubbed audio track and apply the English title/description in the YouTube Studio multi-language UI.