Can I vibecode Repurpose.io?

Build a robust, self-hosted content repurposing and scheduling planner inspired by Repurpose.io, focusing on the workflow organization and scheduling aspect without attempting to reverse-engineer closed platform APIs.
Use exactly this stack: Next.js 15 (App Router), TypeScript, PostgreSQL, Drizzle ORM, BullMQ (for background job processing), and Redis.

### Data Model & Schema
Implement a structured PostgreSQL schema using Drizzle:
- `users`: id, email, name, createdAt.
- `platforms`: id, name (e.g., TikTok, Instagram, YouTube), icon, type (source vs destination), maxVideoLength.
- `connections`: id, userId, platformId, accountName, status, token (encrypted), createdAt.
- `workflows`: id, userId, name, sourceConnectionId, destinationConnectionId, autoPublish (boolean), rules (JSON: e.g., 'only specific hashtags').
- `content_items`: id, workflowId, sourceId (external ID), title, description, mediaUrl, status (pending, scheduled, published, failed), scheduledAt, publishedAt.
- `transformations`: id, contentItemId, removeWatermark (boolean), appendedText, customThumbnailUrl.

### Screens and UX Patterns
1. **Connections Dashboard**: A grid displaying connected accounts (Source and Destination platforms) with 'Connect' and 'Refresh' actions.
2. **Workflow Builder**: A wizard interface allowing users to map a source (e.g., TikTok) to a destination (e.g., YouTube Shorts), set rules (filter by '#repurpose'), and choose manual or auto-publish modes.
3. **Content Queue/Kanban**: A drag-and-drop board or list view showing content in stages: 'Draft', 'Scheduled', 'Published', 'Failed'.
4. **Preview & Edit Modal**: Clicking a draft opens a modal with a side-by-side preview of how the post will look on the destination platform (simulated UI), allowing edits to the caption (e.g., stripping TikTok hashtags and adding YouTube ones).
5. **Export/Download View**: For platforms without open APIs, provide a streamlined "Download & Copy" interface to manually complete the workflow.

### Step-by-Step User Journeys
1. **Setup**: User defines a Workflow from "Google Drive (Source)" to "Instagram Reels (Destination)" set to Manual publish.
2. **Ingestion**: BullMQ background job polls Google Drive every hour, finding new video files and creating `content_items` in the 'Draft' state.
3. **Curation**: User logs in, reviews the Drafts, edits the captions in the Preview Modal to add IG-specific hashtags, and schedules them for specific dates.
4. **Processing**: At the scheduled time, a BullMQ worker picks up the job. Since we are not building proprietary water-mark removal or closed API integrations, the system changes status to "Ready for Manual Export" and sends a browser notification, OR attempts a webhook/API call if an open integration is configured.

### Technical Specifics
- Rely on BullMQ and Redis for robust queue management, retries, and scheduling of ingestion/publishing tasks.
- Keep the app useful as a personal planner even without working external APIs; explain the degraded "manual mode".
- Provide clear empty, loading, validation, success, and failure states.
- Put every secret and external credential in `.env` and provide `.env.example`.
- Add local export functionalities so users are not trapped in the app.
- Write focused unit tests for the data model and the workflow state machine.
- Provide a README with setup instructions, architecture of the background workers, and instructions for running a production-style local build.