diff --git a/.env.example b/.env.example
index b2ec8a9..faecbdf 100644
--- a/.env.example
+++ b/.env.example
@@ -4,3 +4,4 @@ COMPASS_DOCS_PATH=./data/docs
 COMPASS_WORKSPACE=./data/index
 COMPASS_ALLOWED_ORIGINS=http://localhost:3000,http://localhost:5173
 COMPASS_API_KEY=                        # leave empty to disable auth (local dev)
+COMPASS_RATE_LIMIT=60/minute            # requests per IP per minute — raise for trusted networks, lower for public exposure
diff --git a/backend/main.py b/backend/main.py
index 4b5d6a2..f041c52 100644
--- a/backend/main.py
+++ b/backend/main.py
@@ -9,6 +9,9 @@
 from fastapi.security.api_key import APIKeyHeader
 from fastapi.middleware.cors import CORSMiddleware
 from pydantic import BaseModel
+from slowapi import Limiter, _rate_limit_exceeded_handler
+from slowapi.util import get_remote_address
+from slowapi.errors import RateLimitExceeded
 
 load_dotenv()
 
@@ -20,12 +23,29 @@
 logging.basicConfig(level=logging.INFO, format="%(levelname)s — %(name)s — %(message)s")
 logger = logging.getLogger(__name__)
 
-# Config
+# ---------------------------------------------------------------------------
+# Config — all values come from environment variables (see .env.example)
+# ---------------------------------------------------------------------------
+
 OLLAMA_API_BASE = os.getenv("OLLAMA_API_BASE", "http://localhost:11434")
 COMPASS_MODEL   = os.getenv("COMPASS_MODEL", "ollama/gemma4:e4b")
 DOCS_PATH       = Path(os.getenv("COMPASS_DOCS_PATH", "./data/docs"))
 WORKSPACE       = Path(os.getenv("COMPASS_WORKSPACE", "./data/index"))
-API_KEY         = os.getenv("COMPASS_API_KEY")  # None = auth disabled
+
+# Optional API key — if unset, authentication is disabled (safe for local dev).
+# Set COMPASS_API_KEY in .env to require the X-API-Key header on all requests.
+API_KEY = os.getenv("COMPASS_API_KEY")
+
+# Rate limit — all endpoints share the same limit (requests per minute per IP).
+# All requests over the limit receive HTTP 429. Adjust COMPASS_RATE_LIMIT in
+# .env to tune (e.g. "30/minute", "200/minute"). Default: 60/minute.
+RATE_LIMIT = os.getenv("COMPASS_RATE_LIMIT", "60/minute")
+
+os.environ["OLLAMA_API_BASE"] = OLLAMA_API_BASE
+
+# ---------------------------------------------------------------------------
+# Auth
+# ---------------------------------------------------------------------------
 
 _api_key_header = APIKeyHeader(name="X-API-Key", auto_error=False)
 
@@ -34,8 +54,17 @@ async def verify_api_key(key: str = Security(_api_key_header)):
     if API_KEY and key != API_KEY:
         raise HTTPException(status_code=401, detail="Invalid or missing API key.")
 
-os.environ["OLLAMA_API_BASE"] = OLLAMA_API_BASE
 
+# ---------------------------------------------------------------------------
+# Rate limiter — keyed by client IP
+# ---------------------------------------------------------------------------
+
+limiter = Limiter(key_func=get_remote_address)
+
+
+# ---------------------------------------------------------------------------
+# App lifecycle
+# ---------------------------------------------------------------------------
 
 def get_indexer(request: Request) -> CompassIndexer:
     return request.app.state.indexer
@@ -61,6 +90,10 @@ async def lifespan(app: FastAPI):
         app.state.watcher.join()
 
 
+# ---------------------------------------------------------------------------
+# App
+# ---------------------------------------------------------------------------
+
 app = FastAPI(
     title="Compass",
     description="El cerebro operativo de tu empresa.",
@@ -69,6 +102,9 @@ async def lifespan(app: FastAPI):
     dependencies=[Security(verify_api_key)],
 )
 
+app.state.limiter = limiter
+app.add_exception_handler(RateLimitExceeded, _rate_limit_exceeded_handler)
+
 ALLOWED_ORIGINS = [
     o.strip()
     for o in os.getenv("COMPASS_ALLOWED_ORIGINS", "http://localhost:3000,http://localhost:5173").split(",")
@@ -79,16 +115,25 @@ async def lifespan(app: FastAPI):
     CORSMiddleware,
     allow_origins=ALLOWED_ORIGINS,
     allow_methods=["GET", "POST"],
-    allow_headers=["Content-Type"],
+    allow_headers=["Content-Type", "X-API-Key"],
 )
 
 
+# ---------------------------------------------------------------------------
+# Models
+# ---------------------------------------------------------------------------
+
 class ChatRequest(BaseModel):
     question: str
     session_id: str = "default"
 
 
+# ---------------------------------------------------------------------------
+# Routes
+# ---------------------------------------------------------------------------
+
 @app.get("/health")
+@limiter.limit(RATE_LIMIT)
 def health(request: Request):
     indexer = get_indexer(request)
     return {
@@ -99,6 +144,7 @@ def health(request: Request):
 
 
 @app.post("/chat")
+@limiter.limit(RATE_LIMIT)
 async def chat(body: ChatRequest, request: Request):
     indexer = get_indexer(request)
     if not indexer.documents:
@@ -118,6 +164,7 @@ async def chat(body: ChatRequest, request: Request):
 
 
 @app.post("/upload", status_code=201)
+@limiter.limit(RATE_LIMIT)
 async def upload(file: UploadFile = File(...), request: Request = None):
     indexer = get_indexer(request)
 
@@ -139,6 +186,7 @@ async def upload(file: UploadFile = File(...), request: Request = None):
 
 
 @app.get("/documents")
+@limiter.limit(RATE_LIMIT)
 def list_documents(request: Request):
     indexer = get_indexer(request)
     return {"documents": indexer.list_documents(), "total": len(indexer.documents)}
diff --git a/requirements.txt b/requirements.txt
index e17ab4c..beefb59 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -7,6 +7,7 @@ pymupdf==1.26.4
 watchdog==6.0.0
 PyPDF2==3.0.1
 pyyaml==6.0.2
+slowapi==0.1.9
 
 # testing
 pytest==8.3.5
diff --git a/tests/test_api.py b/tests/test_api.py
index 4db0edf..e4be0f3 100644
--- a/tests/test_api.py
+++ b/tests/test_api.py
@@ -50,6 +50,23 @@ def test_auth_applies_to_documents(self, client_empty):
         assert r.status_code == 401
 
 
+# ---------------------------------------------------------------------------
+# Rate limiting
+# ---------------------------------------------------------------------------
+
+class TestRateLimit:
+    def test_rate_limit_handler_is_registered(self, client_empty):
+        """Verify slowapi's 429 handler is registered on the app."""
+        from slowapi.errors import RateLimitExceeded
+        from backend.main import app
+        assert RateLimitExceeded in app.exception_handlers
+
+    def test_limiter_attached_to_app(self, client_empty):
+        """Verify the limiter is attached to app.state so slowapi can find it."""
+        from backend.main import app, limiter
+        assert app.state.limiter is limiter
+
+
 # ---------------------------------------------------------------------------
 # GET /health
 # ---------------------------------------------------------------------------