🧠 NeuroWiki — Your Personal Memory Agent

"The goal isn't to remember everything. It's to never lose what matters."

NeuroWiki is an AI-powered, private knowledge base that transforms anything you read into structured, searchable, interconnected wiki pages — all scoped exclusively to your account.

---

✨ Features

Feature	Description
🔐 Private-by-design	Every user's data lives in a completely isolated tenant. User A cannot see User B's data — ever.
🤖 AI ingestion	Paste a URL, upload a PDF/DOCX/TXT, or type raw text. The AI extracts key knowledge and writes structured wiki pages automatically.
🕸 Knowledge graph	Pages link to each other via [[wikilinks]]. Explore your knowledge visually in the Graph view.
🔍 Semantic search	Ask natural-language questions across your entire knowledge base. Answers are grounded only in your own content.
📓 Diary	Write and store personal diary entries, separate from wiki pages.
🗑 Delete your data	Hover any wiki card → red trash icon appears → confirm → permanently deleted.
📤 Export	Download a full export of your wiki at any time.

---

🚀 Getting Started

1. Prerequisites

• Node.js 20+
• Google Cloud CLI (for deployment only)

2. Clone the repository

git clone https://github.com/SyedArmanAli2003/NeuroWiKi.git
cd NeuroWiKi

3. Install dependencies

npm install

4. Configure environment variables

Copy the example file and fill in your keys:

cp .env.example .env.local

Edit .env.local:

# HydraDB — knowledge storage & semantic search
HYDRADB_API_KEY=sk_live_...

# Google Gemini — AI ingestion
GOOGLE_GENERATIVE_AI_API_KEY=AIza...

# NVIDIA LLaMA (fallback LLM)
NVIDIA_API_KEY=nvapi-...
NVIDIA_MODEL=meta/llama-3.3-70b-instruct

# NextAuth — session encryption
NEXTAUTH_SECRET=your-32-char-random-secret
NEXTAUTH_URL=http://localhost:3000

# Set to 0 to disable AI enrichment on ingest (faster, cheaper)
INGEST_ENRICH=0

⚠️ Never commit .env.local — it is gitignored by default.

5. Run locally

npm run dev

Visit http://localhost:3000. You'll be shown the landing page. After 5 seconds, a sign-in prompt will appear.

---

🏗 Architecture

NeuroWiki
├── app/                     # Next.js App Router pages
│   ├── page.tsx             # Landing page (public, 5s auth prompt)
│   ├── auth/signin          # Sign-in page
│   ├── auth/signup          # Sign-up page
│   ├── wiki/                # Wiki browser + [slug] detail
│   ├── search/              # Semantic search
│   ├── graph/               # Knowledge graph visualisation
│   ├── diary/               # Diary entries
│   ├── ingest/              # Add sources (URL / file / text)
│   └── api/                 # REST API routes
│       ├── auth/            # NextAuth sign-in / sign-up
│       ├── wiki/            # List & delete wiki pages
│       ├── ingest/          # AI ingestion pipeline
│       ├── query/           # Semantic Q&A
│       └── ...
├── lib/
│   ├── auth-options.ts      # Shared NextAuth config
│   ├── user-db.ts           # SQLite users table (bcrypt passwords)
│   ├── db.ts                # SQLite singleton (sources, logs, pages)
│   ├── hydra.ts             # HydraDB client
│   └── hydra-fetch.ts       # HydraDB list/fetch helpers
├── components/
│   └── layout/Topbar.tsx    # Navbar with auth pill
├── middleware.ts            # Route protection (public: /, /auth/*)
└── Dockerfile               # Production container

---

🔐 Authentication

NeuroWiki uses NextAuth.js with a CredentialsProvider backed by a local SQLite database.

• Passwords are hashed with bcrypt (cost factor 12) — never stored in plain text.
• Sessions use JWT stored in an HTTP-only cookie.
• All protected routes are guarded by middleware.ts — unauthenticated requests are redirected to /auth/signin.

Sign Up

1. Go to /auth/signup
2. Enter your name, email, and a password (min 8 characters)
3. A password strength meter gives live feedback
4. On success you're automatically signed in and redirected to the dashboard

Sign In

1. Go to /auth/signin
2. Enter your registered email and password
3. On success you're redirected to the homepage

---

🛡 Data Isolation

Each user gets their own HydraDB tenant (user-{userId}). This means:

• ✅ User A's ingested sources, wiki pages, and diary entries are invisible to User B
• ✅ AI search only queries the current user's knowledge base
• ✅ The delete endpoint validates the requesting user's ID before deleting
• ✅ The wiki list API returns only the current user's pages

---

🗑 Deleting Wiki Entries

On the /wiki page, hover over any knowledge card — a red trash icon appears in the top-right corner of the card. Click it, confirm the prompt, and the entry is permanently deleted from both HydraDB and the local SQLite metadata store.

---

📦 Deployment (Google Cloud Run)

gcloud run deploy neurowiki \
  --source . \
  --project YOUR_PROJECT_ID \
  --region us-central1 \
  --allow-unauthenticated \
  --set-env-vars="NEXTAUTH_URL=https://YOUR_URL,NEXTAUTH_SECRET=...,HYDRADB_API_KEY=...,GOOGLE_GENERATIVE_AI_API_KEY=...,NVIDIA_API_KEY=...,NVIDIA_MODEL=meta/llama-3.3-70b-instruct,INGEST_ENRICH=0"

⚠️ Cloud Run ephemeral disk warning: SQLite files are wiped when a container restarts. This affects the local users table. For a persistent production deployment, migrate the user store to a managed database (e.g., Supabase PostgreSQL, Google Cloud SQL, or Firebase Auth).

---

🌐 Live Demo

https://neurowiki-483008448755.us-central1.run.app

---

📋 Environment Variables Reference

Variable	Required	Description
NEXTAUTH_SECRET	✅	32+ char random string for JWT signing
NEXTAUTH_URL	✅	Full URL of your deployment
HYDRADB_API_KEY	✅	HydraDB API key for knowledge storage
GOOGLE_GENERATIVE_AI_API_KEY	✅	Google Gemini API key for AI ingestion
NVIDIA_API_KEY	✅	NVIDIA API key (fallback LLM)
NVIDIA_MODEL	✅	NVIDIA model name
INGEST_ENRICH	❌	Set to 0 to skip enrichment (recommended)

---

🤝 Contributing

1. Fork the repo
2. Create a feature branch: git checkout -b feature/your-feature
3. Commit your changes
4. Push and open a Pull Request against main

---

📄 License

MIT © 2024 NeuroWiki. Your knowledge. Your control.