diff --git a/.env.example b/.env.example
index b7a9e22..8656cef 100644
--- a/.env.example
+++ b/.env.example
@@ -34,5 +34,8 @@ JWT_SECRET=change-this-in-production-min-32-bytes
 JWT_ALGORITHM=HS256
 ACCESS_TOKEN_MINUTES=720
 
+# ── CORS ───────────────────────────────────────────────────────
+CORS_ALLOW_ORIGINS=http://localhost:5173,http://127.0.0.1:5173,http://localhost:3000,http://127.0.0.1:3000
+
 # ── Database ───────────────────────────────────────────────────
 DATABASE_URL=sqlite:///./assistant.db
diff --git a/README.md b/README.md
index d38e8e5..e0c2280 100644
--- a/README.md
+++ b/README.md
@@ -118,6 +118,7 @@ Important variables:
 - `JWT_SECRET`
 - `DATABASE_URL`
 - `RATE_LIMIT_PER_MINUTE`
+- `CORS_ALLOW_ORIGINS`
 - `LLM_API_KEY` (optional)
 
 The app can still run without external AI providers when `LLM_ENABLED=false`.
@@ -372,6 +373,7 @@ The backend includes built-in resilience for LLM requests:
 | Variable | Default | Description |
 |---|---|---|
 | `RATE_LIMIT_PER_MINUTE` | `30` | Max requests per IP per minute |
+| `CORS_ALLOW_ORIGINS` | Local dev + Render app origins | Comma-separated browser origins allowed to call the API |
 | `LLM_ENABLED` | `false` | Enable LLM provider |
 | `LLM_API_KEY` | — | API key for your LLM provider |
 | `LLM_BASE_URL` | `https://api.openai.com/v1` | LLM base URL |
diff --git a/backend/app/config.py b/backend/app/config.py
index 06146d1..6908777 100644
--- a/backend/app/config.py
+++ b/backend/app/config.py
@@ -41,6 +41,15 @@ def _bool_env(name: str, default: bool) -> bool:
     return raw_value.strip().lower() in {"1", "true", "yes", "on"}
 
 
+def _csv_env(name: str, default: list[str]) -> list[str]:
+    raw_value = os.getenv(name)
+    if raw_value is None:
+        return default
+
+    values = [value.strip() for value in raw_value.split(",") if value.strip()]
+    return values or default
+
+
 class Settings:
     """Application settings loaded from environment variables."""
 
@@ -62,6 +71,16 @@ class Settings:
     jwt_secret: str = os.getenv("JWT_SECRET", "change-this-in-production-min-32-bytes")
     jwt_algorithm: str = os.getenv("JWT_ALGORITHM", "HS256")
     access_token_minutes: int = _int_env("ACCESS_TOKEN_MINUTES", 720)
+    cors_allow_origins: list[str] = _csv_env(
+        "CORS_ALLOW_ORIGINS",
+        [
+            "http://localhost:5173",
+            "http://127.0.0.1:5173",
+            "http://localhost:3000",
+            "http://127.0.0.1:3000",
+            "https://qyverixai.onrender.com",
+        ],
+    )
     llm_enabled: bool = _bool_env("LLM_ENABLED", False)
     llm_api_key: str | None = os.getenv("LLM_API_KEY")
     llm_base_url: str = os.getenv("LLM_BASE_URL", "https://api.openai.com/v1")
diff --git a/backend/app/main.py b/backend/app/main.py
index e868905..dc1562f 100644
--- a/backend/app/main.py
+++ b/backend/app/main.py
@@ -30,6 +30,7 @@
 from .services import database
 from .services.scheduler import start_scheduler, stop_scheduler
 from .database import Base, engine
+from .config import settings
 
 from .schemas import HealthResponse
 
@@ -85,10 +86,18 @@ async def lifespan(app: FastAPI):
 
 # ── Middleware ────────────────────────────────────────────────────────────────
 app.add_middleware(GZipMiddleware, minimum_size=1000)
+
+
+def _allow_credentials_for_origins(origins: list[str]) -> bool:
+    return "*" not in origins
+
+
+_cors_allow_origins = settings.cors_allow_origins
+_cors_allow_credentials = _allow_credentials_for_origins(_cors_allow_origins)
 app.add_middleware(
     CORSMiddleware,
-    allow_origins=["*"],
-    allow_credentials=True,
+    allow_origins=_cors_allow_origins,
+    allow_credentials=_cors_allow_credentials,
     allow_methods=["*"],
     allow_headers=["*"],
 )
diff --git a/backend/tests/test_endpoints.py b/backend/tests/test_endpoints.py
index c4d776b..a622afc 100644
--- a/backend/tests/test_endpoints.py
+++ b/backend/tests/test_endpoints.py
@@ -188,6 +188,40 @@ def test_rate_limit_returns_429_with_retry_after_header():
     assert r.headers["X-RateLimit-Remaining"] == "0"
 
 
+def test_cors_allows_configured_origin_with_credentials():
+    r = client.options(
+        "/auth/me",
+        headers={
+            "Origin": "http://localhost:5173",
+            "Access-Control-Request-Method": "GET",
+            "Access-Control-Request-Headers": "authorization",
+        },
+    )
+
+    assert r.status_code == 200
+    assert r.headers["access-control-allow-origin"] == "http://localhost:5173"
+    assert r.headers["access-control-allow-credentials"] == "true"
+
+
+def test_cors_rejects_unconfigured_origin():
+    r = client.options(
+        "/auth/me",
+        headers={
+            "Origin": "https://attacker.example",
+            "Access-Control-Request-Method": "GET",
+            "Access-Control-Request-Headers": "authorization",
+        },
+    )
+
+    assert r.status_code == 400
+    assert "access-control-allow-origin" not in r.headers
+
+
+def test_wildcard_cors_disables_credentials():
+    assert app_main._allow_credentials_for_origins(["*"]) is False
+    assert app_main._allow_credentials_for_origins(["http://localhost:5173"]) is True
+
+
 # ── Explanation ───────────────────────────────────────────────────────────────
 def test_explanation_python():
     r = client.post("/explanation/", json={"code": PYTHON_CLEAN, "language": "python"})
diff --git a/render.yaml b/render.yaml
index a28b9fe..e9d06a5 100644
--- a/render.yaml
+++ b/render.yaml
@@ -10,8 +10,10 @@ services:
         value: 30
       - key: LLM_ENABLED
         value: false
+      - key: CORS_ALLOW_ORIGINS
+        value: https://qyverixai.onrender.com
     staticPublishPath: frontend
     routes:
       - type: rewrite
         source: /app
-        destination: /frontend/index.html
\ No newline at end of file
+        destination: /frontend/index.html