Build an AI faceless YouTube video generator web app with these features:

1. Topic input: user enters a niche and topic (or picks from trending topics pulled from Google Trends or Reddit for the niche), video length (short, standard, long), and tone (calm documentary, energetic listicle, cozy facts). The app maintains saved channel profiles with recurring series formats.

2. AI script writer: generate a script with a hook in the first 15 seconds, chaptered body sections, pattern interrupts every 30 to 45 seconds, and a subscribe CTA at the end. User can edit the script inline before rendering.

3. AI voiceover: TTS with 8 to 12 voice options across accents and styles, speed control, and per-line pronunciation overrides. Auto-detect and fix awkward phrasing before rendering.

4. Visual engine: for each script section, generate or fetch visuals. Modes: AI-generated images (prompted per scene), stock footage search by keyword, and simple motion effects (Ken Burns zoom, pan) to keep still images alive. Auto match visual duration to narration.

5. Subtitles: auto-generate burned-in subtitles in a customizable style (font, color, position) synced word by word for the retention-boosting caption look.

6. Music and SFX: background music library by mood with auto-ducking under narration, plus optional whoosh and impact effects at section transitions.

7. Render pipeline: produce a 1080p MP4 with a render queue, progress bar, and downloadable output. Support Shorts (9:16) and standard (16:9) formats.

8. Batch mode: queue a week of videos from a list of topics and render overnight.

9. Originality and policy guardrails: script plagiarism check against top-ranking videos on the same topic, readability scoring, and a pre-publish checklist warning about YouTube's repetitive-content and disclosure policies for synthetic media, including a reminder to review and edit before uploading.

10. Title and description pack: generate 5 title options, description with timestamps, tags, and a pinned-comment idea to go with the video.