Can I vibe code PipiAds?

Build a internal TikTok Ad Vault web application using Next.js (App Router), Tailwind CSS, Supabase, and Meilisearch.

1. DATA MODEL:
- Create a 'saved_ads' table in Supabase: id (uuid), ad_title (text), brand_name (text), landing_page_url (text), video_url (text), thumbnail_url (text), views (int), likes (int), country (text), niche (text), created_at (timestamp).
- Create a 'collections' table to organize saved ads into folders.

2. CORE FUNCTIONALITY:
- Public-facing Dashboard: Display indexed ads in a responsive CSS grid of cards. Each card must render a video preview, brand name, estimated engagement metrics, and tags.
- Search & Filters: Connect Meilisearch to provide instant instant-search filtering by keyword, country, engagement threshold, and target niche.
- Manual Ad Importer: Provide an API endpoint (`/api/import-ad`) that accepts a TikTok video URL, uses a headless scraper library to extract the video source file and metadata, uploads the video file to Supabase Storage, and inserts the record into Postgres and Meilisearch.
- Landing Page Inspector: Parse target URLs from incoming ads to detect e-commerce technology stacks (e.g., check for Shopify window objects or WooCommerce markers) and save findings in the ad payload.

3. FAILURE HANDLING:
- If video extraction fails due to bot detection, fallback to storing raw metadata and return an operational status alert to the UI.
- Ensure video playback falls back to a static image thumbnail if the CDN stream breaks.

4. OUT OF SCOPE:
- Do not build continuous multi-million ad crawling pipelines or proxy management systems. Rely solely on manual submission or API webhook ingestion.

$ each button prefixes agent-specific run instructions · build your own product, never copy proprietary code, trademarks or designs