Wandrmark — Smart Local Explorer & Travel Planner

Quick Start (with Docker)

Docker runs Redis for you. AI is powered by NVIDIA NIM (cloud API — no local model needed).

# 1. Clone & install
git clone https://github.com/divarun/wandrmark.git
cd wandrmark
npm install

# 2. Configure environment
cp .env.example .env
# Edit .env — set NVIDIA_API_KEY (required for AI features)
# Get a free key at https://build.nvidia.com/

# 3. Copy .env into backend
cp .env backend/.env

# 4. Start infrastructure (Redis)
docker compose up -d

# 5. Start everything
npm run dev
# → Frontend: http://localhost:3000
# → Backend:  http://localhost:3001

Quick Start (without Docker)

You need Redis running locally (or use a cloud Redis service). Redis is optional — the app works without it, just without caching.

# 1. Clone & install
git clone https://github.com/divarun/wandrmark.git
cd wandrmark
npm install

# 2. Configure environment
cp .env.example .env
# Edit .env with your Redis connection string and NVIDIA_API_KEY

# 3. Copy .env into backend
cp .env backend/.env

# 4. Start everything
npm run dev
# → Frontend: http://localhost:3000
# → Backend:  http://localhost:3001

Note: Redis is optional. The app works without it — caching and cache warming are simply disabled.

Environment Variables

Copy .env.example to .env and backend/.env. Key variables:

Variable	Required	Description
NVIDIA_API_KEY	Yes (for AI)	NIM API key from build.nvidia.com
REDIS_URL	No	Redis URL — defaults to redis://localhost:6379
NIM_MODEL	No	Model name — defaults to meta/llama-3.1-8b-instruct

Optional

Pre-warm the cache

The backend auto-warms on startup if Redis is empty. You can also trigger it manually:

# Warm everything (geocoding + POI data + AI city insights)
cd backend
npm run warm-cache

# Or just geocoding
npm run warm-geocoding

Cache warming pre-generates AI insights for all major cities so users see instant results on first load. Insights are cached for 7 days; the cache refreshes automatically every Sunday at midnight.

Redis Insights

1. Open browser: http://localhost:5540 (requires Docker)

2. Add a Standalone Redis database:

   Field	Value
   Host	wandrmark-redis
   Port	6379
   Password	(leave blank)
   Database	0

3. Test connection → should succeed ✅

4. Browse keys and view data immediately.

See Full Developer Guide for more details