diff --git a/VERSION b/VERSION
index b1e80bb..845639e 100644
--- a/VERSION
+++ b/VERSION
@@ -1 +1 @@
-0.1.3
+0.1.4
diff --git a/alembic/env.py b/alembic/env.py
index 81ae248..12d8a51 100644
--- a/alembic/env.py
+++ b/alembic/env.py
@@ -15,6 +15,7 @@
 
 # Import models so Base.metadata has all table definitions registered
 import lenny.core.models  # noqa: F401
+import lenny.core.cache  # noqa: F401
 
 # Alembic Config object — access to alembic.ini values
 config = context.config
diff --git a/alembic/versions/c6b7da6debc2_add_cache_table.py b/alembic/versions/c6b7da6debc2_add_cache_table.py
new file mode 100644
index 0000000..83d56d1
--- /dev/null
+++ b/alembic/versions/c6b7da6debc2_add_cache_table.py
@@ -0,0 +1,43 @@
+"""add cache table
+
+Revision ID: c6b7da6debc2
+Revises: 001_baseline
+Create Date: 2026-03-28 18:58:54.334094
+
+"""
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = 'c6b7da6debc2'
+down_revision: Union[str, None] = '001_baseline'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.create_table('cache',
+    sa.Column('id', sa.BigInteger(), nullable=False),
+    sa.Column('scope', sa.String(length=64), nullable=False),
+    sa.Column('key', sa.String(length=255), nullable=False),
+    sa.Column('value', sa.String(length=1024), nullable=True),
+    sa.Column('created_at', sa.DateTime(timezone=True), server_default=sa.text('now()'), nullable=True),
+    sa.Column('expires_at', sa.DateTime(timezone=True), nullable=False),
+    sa.PrimaryKeyConstraint('id'),
+    prefixes=['UNLOGGED']
+    )
+    op.create_index('idx_cache_expires', 'cache', ['expires_at'], unique=False)
+    op.create_index('idx_cache_scope_key_expires', 'cache', ['scope', 'key', 'expires_at'], unique=False)
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_index('idx_cache_scope_key_expires', table_name='cache')
+    op.drop_index('idx_cache_expires', table_name='cache')
+    op.drop_table('cache')
+    # ### end Alembic commands ###
diff --git a/compose.yaml b/compose.yaml
index e6df61a..4289c00 100644
--- a/compose.yaml
+++ b/compose.yaml
@@ -13,7 +13,7 @@ services:
       s3:
         condition: service_healthy
     healthcheck:
-      test: ["CMD-SHELL", "curl -sf http://localhost:1337/v1/api/opds || exit 1"]
+      test: ["CMD-SHELL", "curl -sf http://localhost:1337/v1/api/health || exit 1"]
       interval: 10s
       timeout: 5s
       retries: 3
diff --git a/docker/api/Dockerfile b/docker/api/Dockerfile
index 31ab113..1b602e7 100644
--- a/docker/api/Dockerfile
+++ b/docker/api/Dockerfile
@@ -32,4 +32,4 @@ COPY ./docker/nginx/conf.d/lenny.conf /etc/nginx/conf.d/lenny.conf
 COPY ./docker/utils/migrate.sh /app/docker/utils/migrate.sh
 RUN chmod +x /app/docker/utils/migrate.sh
 
-CMD ["sh", "-c", "sh /app/docker/utils/migrate.sh && python -m uvicorn lenny.app:app --host 0.0.0.0 --port 1337 --workers=${LENNY_WORKERS:-1} --log-level=${LENNY_LOG_LEVEL:-info} $([ \"${LENNY_PRODUCTION:-true}\" = \"false\" ] && echo --reload) & exec nginx"]
+CMD ["sh", "-c", "sh /app/docker/utils/migrate.sh && python -m uvicorn lenny.app:app --host 0.0.0.0 --port 1337 --workers=${LENNY_WORKERS:-$([ \"${LENNY_PRODUCTION:-true}\" = \"false\" ] && echo 1 || echo 3)} --log-level=${LENNY_LOG_LEVEL:-info} $([ \"${LENNY_PRODUCTION:-true}\" = \"false\" ] && echo --reload) & exec nginx"]
diff --git a/docker/configure.sh b/docker/configure.sh
index 05a2e5a..89fbf73 100755
--- a/docker/configure.sh
+++ b/docker/configure.sh
@@ -17,7 +17,7 @@ else
   # Use environment variables if they are set, otherwise provide defaults or generate secure values
   LENNY_HOST="localhost"
   LENNY_PORT="${LENNY_PORT:-8080}"
-  LENNY_WORKERS="${LENNY_WORKERS:-1}"
+  LENNY_WORKERS="${LENNY_WORKERS:-3}"
   LENNY_LOG_LEVEL="${LENNY_LOG_LEVEL:-debug}"
   LENNY_PRODUCTION="${LENNY_PRODUCTION:-true}"
   LENNY_SSL_CRT="${LENNY_SSL_CRT:-}"
@@ -96,3 +96,27 @@ NODE_ENV=$NODE_ENV
 
 EOF
 fi
+
+# Install 'lenny' CLI command if not already available
+LENNY_PROJECT_DIR="$(cd "$(dirname "$0")/.." && pwd)"
+if ! command -v lenny &>/dev/null; then
+  INSTALL_DIR="$HOME/.local/bin"
+  mkdir -p "$INSTALL_DIR"
+  cat > "$INSTALL_DIR/lenny" <<SCRIPT
+#!/bin/sh
+make -C "$LENNY_PROJECT_DIR" "\$@"
+SCRIPT
+  chmod +x "$INSTALL_DIR/lenny"
+
+  case ":$PATH:" in
+    *":$INSTALL_DIR:"*)
+      echo "[lenny] CLI installed. You can now use: lenny start, lenny stop, etc."
+      ;;
+    *)
+      echo "[lenny] CLI installed to $INSTALL_DIR/lenny"
+      echo "[lenny] Add to PATH: echo 'export PATH=\"\$HOME/.local/bin:\$PATH\"' >> ~/.zshrc && source ~/.zshrc"
+      ;;
+  esac
+else
+  echo "Skipping CLI install: 'lenny' command already available."
+fi
diff --git a/docker/nginx/conf.d/lenny.conf b/docker/nginx/conf.d/lenny.conf
index fafaaff..d855296 100644
--- a/docker/nginx/conf.d/lenny.conf
+++ b/docker/nginx/conf.d/lenny.conf
@@ -6,7 +6,37 @@ server {
     error_log /var/log/nginx/error.log debug;
     access_log /var/log/nginx/access.log;
 
+    # Sensitive routes: tighter rate limits
+    location ~ ^/v1/api/items/[^/]+/borrow {
+        limit_req zone=borrow burst=3 nodelay;
+        proxy_pass http://lenny_api:1337$request_uri;
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+    }
+
+    location /v1/api/oauth/authorize {
+        limit_req zone=oauth burst=3 nodelay;
+        proxy_pass http://lenny_api:1337/v1/api/oauth/authorize;
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+    }
+
+    location = /v1/api/upload {
+        limit_req zone=upload burst=2 nodelay;
+        proxy_pass http://lenny_api:1337/v1/api/upload;
+        proxy_set_header Host $host;
+        proxy_set_header X-Real-IP $remote_addr;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_set_header X-Forwarded-Proto $scheme;
+    }
+
+    # General API: 30 req/min
     location /v1/api {
+        limit_req zone=api burst=10 nodelay;
         proxy_pass http://lenny_api:1337/v1/api;
         proxy_set_header Host $host;
         proxy_set_header X-Real-IP $remote_addr;
diff --git a/docker/nginx/nginx.conf b/docker/nginx/nginx.conf
index 74e2a88..52e4f93 100644
--- a/docker/nginx/nginx.conf
+++ b/docker/nginx/nginx.conf
@@ -15,6 +15,14 @@ http {
     sendfile        on;
     keepalive_timeout  65;
     client_max_body_size 50M;
+
+    # Rate limiting: shared memory zones, no DB overhead.
+    limit_req_zone $binary_remote_addr zone=api:10m rate=30r/m;
+    limit_req_zone $binary_remote_addr zone=borrow:5m rate=10r/m;
+    limit_req_zone $binary_remote_addr zone=oauth:5m rate=10r/m;
+    limit_req_zone $binary_remote_addr zone=upload:5m rate=5r/m;
+    limit_req_status 429;
+
     include /etc/nginx/conf.d/*.conf;
 }
 
diff --git a/lenny/configs/__init__.py b/lenny/configs/__init__.py
index 7e04fc6..b5495ee 100644
--- a/lenny/configs/__init__.py
+++ b/lenny/configs/__init__.py
@@ -18,7 +18,7 @@
 PROXY = os.environ.get('LENNY_PROXY', '')
 HOST = os.environ.get('LENNY_HOST', 'localhost')
 PORT = int(os.environ.get('LENNY_PORT', 8080))
-WORKERS = int(os.environ.get('LENNY_WORKERS', 1))
+WORKERS = int(os.environ.get('LENNY_WORKERS', 1 if TESTING else 3))
 DEBUG = bool(int(os.environ.get('LENNY_DEBUG', 0)))
 SEED = os.environ.get('LENNY_SEED')
 LOG_LEVEL = os.environ.get('LENNY_LOG_LEVEL', 'info')
diff --git a/lenny/core/auth.py b/lenny/core/auth.py
index d9578d8..a59b55f 100644
--- a/lenny/core/auth.py
+++ b/lenny/core/auth.py
@@ -1,12 +1,11 @@
 import hashlib
-import hmac
 import logging
-import time
 import httpx
 from datetime import datetime, timedelta
 from typing import Optional
 from itsdangerous import URLSafeTimedSerializer, BadSignature
 from lenny.configs import SEED, OTP_SERVER
+from lenny.core.cache import Cache
 from lenny.core.exceptions import RateLimitError
 
 logging.basicConfig(
@@ -83,9 +82,6 @@ def verify_session_cookie(session, client_ip: str = None):
         
 class OTP:
 
-    _attempts = {}
-    _send_attempts = {}
-
     @classmethod
     def generate(cls, email: str, issued_minute: int = None) -> str:
         """
@@ -112,11 +108,9 @@ def verify(cls, email: str, ip_address: str, otp: str) -> bool:
     @classmethod
     def is_send_rate_limited(cls, email: str) -> bool:
         """Limit OTP send requests: 5 emails per 5 minutes per email."""
-        now = time.time()
-        attempts = cls._send_attempts.get(email, [])
-        attempts = [ts for ts in attempts if now - ts < EMAIL_WINDOW_SECONDS]
-        cls._send_attempts[email] = attempts + [now]
-        return len(attempts) >= EMAIL_REQUEST_LIMIT
+        return Cache.is_throttled(
+            "otp:send", email, EMAIL_REQUEST_LIMIT, EMAIL_WINDOW_SECONDS
+        )
 
     @classmethod
     def issue(cls, email: str, ip_address: str) -> dict:
@@ -139,15 +133,10 @@ def redeem(cls, email: str, ip_address: str, otp: str) -> bool:
 
     @classmethod
     def is_rate_limited(cls, email: str) -> bool:
-        """Updates attempts within timeframe for email and
-        returns True if the user is making too many attempts.
-        """
-        now = time.time()
-        attempts = cls._attempts.get(email, [])
-        # Keep only recent attempts
-        attempts = [ts for ts in attempts if now - ts < ATTEMPT_WINDOW_SECONDS]
-        cls._attempts[email] = attempts + [now]
-        return len(attempts) >= ATTEMPT_LIMIT
+        """Returns True if the user is making too many OTP verification attempts."""
+        return Cache.is_throttled(
+            "otp:verify", email, ATTEMPT_LIMIT, ATTEMPT_WINDOW_SECONDS
+        )
 
     @classmethod
     def authenticate(cls, email: str, otp: str, ip: str = None) -> Optional[str]:
diff --git a/lenny/core/cache.py b/lenny/core/cache.py
new file mode 100644
index 0000000..5e3b4e7
--- /dev/null
+++ b/lenny/core/cache.py
@@ -0,0 +1,118 @@
+"""
+    PostgreSQL-backed cache for Lenny.
+
+    Used for OTP email-based rate limiting across multiple Uvicorn workers.
+    IP-based rate limiting is handled by nginx (limit_req).
+
+    :copyright: (c) 2015 by AUTHORS
+    :license: see LICENSE for more details
+"""
+
+import logging
+import random
+from datetime import datetime, timedelta, timezone
+
+from sqlalchemy import Column, String, BigInteger, DateTime, Index
+from sqlalchemy.sql import func
+
+from lenny.core.db import session as db, Base
+from lenny.core.exceptions import DatabaseInsertError
+from lenny import configs
+
+logger = logging.getLogger(__name__)
+
+PURGE_PROBABILITY = 0.01  # 1-in-100 chance per rate limit check
+
+# UNLOGGED tables skip WAL for faster writes — ideal for ephemeral cache.
+# SQLite (used in tests) doesn't support UNLOGGED, so we only apply it on PostgreSQL.
+_cache_table_opts = {'prefixes': ['UNLOGGED']} if not configs.TESTING else {}
+
+
+class CacheEntry(Base):
+    __tablename__ = 'cache'
+    __table_args__ = (
+        Index('idx_cache_scope_key_expires', 'scope', 'key', 'expires_at'),
+        Index('idx_cache_expires', 'expires_at'),
+        _cache_table_opts,
+    )
+
+    id = Column(BigInteger, primary_key=True)
+    scope = Column(String(64), nullable=False)
+    key = Column(String(255), nullable=False)
+    value = Column(String(1024), nullable=True)
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    expires_at = Column(DateTime(timezone=True), nullable=False)
+
+
+class Cache:
+    """PostgreSQL-backed cache with built-in rate limiting."""
+
+    @classmethod
+    def _record(cls, scope, key, ttl, value=None):
+        """Insert a cache entry that expires after ttl seconds."""
+        try:
+            entry = CacheEntry(
+                scope=scope,
+                key=key,
+                value=value,
+                expires_at=datetime.now(timezone.utc) + timedelta(seconds=ttl),
+            )
+            db.add(entry)
+            db.commit()
+            return entry
+        except Exception as e:
+            db.rollback()
+            raise DatabaseInsertError(f"Failed to record cache entry: {str(e)}")
+
+    @classmethod
+    def _count(cls, scope, key):
+        """Count unexpired entries for a given scope and key."""
+        try:
+            now = datetime.now(timezone.utc)
+            count = db.query(CacheEntry).filter(
+                CacheEntry.scope == scope,
+                CacheEntry.key == key,
+                CacheEntry.expires_at > now,
+            ).count()
+            db.rollback()
+            return count
+        except Exception as e:
+            db.rollback()
+            logger.warning(f"Cache count failed: {str(e)}")
+            return 0
+
+    @classmethod
+    def is_throttled(cls, scope, key, limit, ttl):
+        """Check if a key has exceeded its rate limit.
+
+        Counts existing unexpired entries, then records the current
+        attempt. Returns True if count >= limit (before recording).
+        Only records the attempt if not already throttled.
+        """
+        current_count = cls._count(scope, key)
+
+        if current_count < limit:
+            try:
+                cls._record(scope, key, ttl)
+            except DatabaseInsertError:
+                pass
+
+        if random.random() < PURGE_PROBABILITY:
+            cls.purge()
+
+        return current_count >= limit
+
+    @classmethod
+    def purge(cls):
+        """Delete all expired cache entries."""
+        try:
+            now = datetime.now(timezone.utc)
+            deleted = db.query(CacheEntry).filter(
+                CacheEntry.expires_at < now,
+            ).delete()
+            db.commit()
+            if deleted:
+                logger.debug(f"Cache purge: removed {deleted} expired entries")
+        except Exception as e:
+            db.rollback()
+            logger.warning(f"Cache purge failed: {str(e)}")
diff --git a/lenny/routes/api.py b/lenny/routes/api.py
index 2b78c71..ab0fedb 100644
--- a/lenny/routes/api.py
+++ b/lenny/routes/api.py
@@ -90,6 +90,8 @@ def is_direct_auth_mode(auth_mode: Optional[str] = None, beta: bool = False) ->
     return (auth_mode == "direct") or beta or configs.AUTH_MODE_DIRECT
 
 
+# All IP-based rate limiting is handled by nginx (limit_req zones).
+# OTP email-based rate limiting remains in auth.py via Cache.is_throttled.
 router = APIRouter()
 
 def requires_item_auth(do_function=None):
@@ -128,6 +130,10 @@ async def home(request: Request):
     kwargs = {"request": request}
     return request.app.templates.TemplateResponse("index.html", kwargs)
 
+@router.get('/health', status_code=status.HTTP_200_OK)
+async def health():
+    return {"status": "ok"}
+
 @router.get("/items")
 async def get_items(fields: Optional[str]=None, offset: Optional[int]=None, limit: Optional[int]=None):
     fields = fields.split(",") if fields else None