Can I vibecode Aurelius?

Build a personal replacement for Aurelius in an empty repository.
Use Next.js 15 (App Router), TypeScript, Tailwind CSS, PostgreSQL, Drizzle ORM, and OpenAI API for AI summarization; do not offer alternative stacks.

# Core Value Proposition & Scope
The goal is to create a powerful, lightweight UX research repository and synthesis platform. The app bridges the gap between raw research data and actionable product outcomes. It should focus on "powerful simplicity," allowing researchers to transition from raw notes to key insights that directly influence product decisions without unnecessary complexity.

# Key Features & Data Model
1. **Research Projects & Sessions:** Organize research into Projects. Each project contains multiple Sessions (interviews, usability tests, etc.).
2. **Raw Data & Notes Management:** A centralized interface to paste transcripts, capture raw notes, and store user feedback.
3. **Analysis & Tagging Engine:**
   - Allow users to select text snippets within notes/transcripts and assign Tags (e.g., "pain point", "feature request", "positive feedback").
   - Support bulk tagging and pattern finding.
4. **Affinity Board / Synthesis View:**
   - A visual kanban-style board where tags and snippets can be organized into Themes and Groups to make sense of qualitative data.
   - Drag-and-drop interface for affinity mapping.
5. **AI Assist:**
   - Integrate OpenAI to automatically generate summary paragraphs and extract key themes from a session's raw data.
   - AI should suggest tags for untagged notes based on project history.
6. **Insights & Outcomes:**
   - Create "Key Insights" that synthesize multiple themes and snippets.
   - Link Key Insights directly to action items or user stories. Provide a mock integration export (e.g., JSON export mimicking a Jira ticket creation payload).
7. **Reporting:**
   - Automatically generate a "Research Report" view for a Project, summarizing Key Insights, top Themes, and supporting quotes.

# UI/UX & Layout
- Use a modern, clean, minimalist design typical of high-end SaaS tools (think Notion meets Linear).
- **Sidebar:** Navigation for Projects, All Notes, Tags, Insights, and Settings.
- **Main Content Area:** 
  - **Project Dashboard:** Overview of sessions, recent tags, and AI summary.
  - **Note Editor:** Split view with the raw transcript on the left, and a tagging/highlighting interface on the right.
  - **Affinity Board:** Horizontal scrollable columns for Themes, containing cards for each tagged snippet.
- Implement rich micro-interactions (hover states, smooth transitions) and use a curated color palette (e.g., Slate, Indigo, and Emerald).

# Engineering Requirements
- Make the first run work locally with one documented command (`npm run dev`).
- Store all user data locally by default (using a local Postgres instance or SQLite).
- Put secrets in `.env`, ship `.env.example`.
- Include clear empty states, loading skeletons, success toasts, and recoverable error states.
- Add input validation (zod) and graceful handling of unavailable AI APIs.

# Deliberately Excluded
- Do not add user accounts, billing, telemetry, or analytics.
- Deliberately leave out audio/video storage and automated transcription (rely on pasting text).
- Deliberately leave out enterprise SSO and governance.

Finish by running basic unit tests for the data transformation logic and listing the exact setup commands used in the README.