diff --git a/.gitignore b/.gitignore
index a547bf3..a41eaca 100644
--- a/.gitignore
+++ b/.gitignore
@@ -22,3 +22,9 @@ dist-ssr
 *.njsproj
 *.sln
 *.sw?
+
+# Playwright E2E
+/playwright-report/
+/test-results/
+/blob-report/
+/playwright/.cache/
diff --git a/package.json b/package.json
index b7ed7eb..e1766ac 100644
--- a/package.json
+++ b/package.json
@@ -8,7 +8,10 @@
     "build": "vite build",
     "lint": "eslint .",
     "preview": "vite preview",
-    "test": "vitest"
+    "test": "vitest",
+    "test:e2e": "playwright test",
+    "test:e2e:ui": "playwright test --ui",
+    "test:e2e:report": "playwright show-report"
   },
   "engines": {
     "node": ">=20.0.0"
@@ -57,6 +60,7 @@
   },
   "devDependencies": {
     "@eslint/js": "^9.33.0",
+    "@playwright/test": "^1.49.1",
     "@testing-library/dom": "^10.4.1",
     "@testing-library/react": "^16.3.2",
     "@types/react": "^19.1.10",
diff --git a/playwright.config.js b/playwright.config.js
new file mode 100644
index 0000000..07e48cb
--- /dev/null
+++ b/playwright.config.js
@@ -0,0 +1,54 @@
+import { defineConfig, devices } from '@playwright/test';
+
+/**
+ * Playwright E2E configuration for AxeCode frontend.
+ *
+ * Prerequisites (see tests/e2e/README.md for full details):
+ *  - A running backend the app can talk to (local docker-compose, or a
+ *    staging deployment) reachable at E2E_BASE_URL / VITE_API_BASE_URL.
+ *  - The backend and frontend must both be configured with Google's
+ *    public reCAPTCHA TEST key pair (documented in tests/e2e/README.md)
+ *    so the login/register forms can be automated without solving a
+ *    real challenge. Never use the test key pair in production.
+ *  - Seeded test user credentials via E2E_TEST_USER_EMAIL / E2E_TEST_USER_PASSWORD.
+ */
+const BASE_URL = process.env.E2E_BASE_URL || 'http://localhost:5173';
+
+export default defineConfig({
+  testDir: './tests/e2e',
+  outputDir: './test-results',
+  fullyParallel: true,
+  forbidOnly: !!process.env.CI,
+  retries: process.env.CI ? 2 : 0,
+  workers: process.env.CI ? 1 : undefined,
+  reporter: [
+    ['html', { outputFolder: 'playwright-report', open: 'never' }],
+    ['list'],
+  ],
+
+  use: {
+    baseURL: BASE_URL,
+    trace: 'on-first-retry',
+    screenshot: 'only-on-failure',
+    video: 'retain-on-failure',
+  },
+
+  projects: [
+    {
+      name: 'chromium',
+      use: { ...devices['Desktop Chrome'] },
+    },
+  ],
+
+  // Only auto-start the dev server locally. In CI we assume the frontend
+  // (and the backend it talks to) are already running against a staging
+  // environment — see the "Full-Stack E2E" section of the QA plan.
+  webServer: process.env.CI
+    ? undefined
+    : {
+        command: 'npm run dev',
+        url: BASE_URL,
+        reuseExistingServer: true,
+        timeout: 120 * 1000,
+      },
+});
diff --git a/tests/e2e/README.md b/tests/e2e/README.md
new file mode 100644
index 0000000..2a005ac
--- /dev/null
+++ b/tests/e2e/README.md
@@ -0,0 +1,73 @@
+# E2E Tests (Playwright)
+
+This is the first E2E test suite for the AxeCode frontend — previously
+there was no E2E coverage at all, only Vitest unit tests under
+`tests/useCases`. See the QA/scalability plan for the full context.
+
+## What's covered so far
+
+- `auth.spec.js` — login with valid credentials, and the error path for
+  invalid credentials (no false-positive redirect, form re-enabled).
+- `problems.spec.js` — the `/problems` catalog loads while authenticated,
+  and opening a problem renders the workspace (Run/Submit, tabs, Monaco
+  editor mounted).
+- `courses.spec.js` — the `/courses` catalog loads while authenticated,
+  and opening a course renders its details page.
+
+**Not covered yet (intentionally, tracked separately):**
+- Actual Judge0 code execution results (Run/Submit correctness) — needs
+  seeded test problems with known-good solutions in a staging environment.
+- The full purchase flow (checkout → Paymob/Kashier → entitlement granted
+  → content unlocked) — needs sandboxed payment provider credentials.
+- Live/Agora video call flows.
+
+These are P0/P1 items in the QA plan and are meant to build on this setup,
+not block it.
+
+## Why every test logs in through the real UI
+
+Every route except the landing page, auth pages, and legal pages is
+wrapped in `ProtectedRoute` (see `src/presentation/routes/AppRoutes.jsx`),
+so there is no "public" content to test against. Login itself requires
+solving reCAPTCHA — the submit button on `LoginPage` stays disabled until
+`recaptchaToken` is set, independently of the backend's `RECAPTCHA_REQUIRED`
+flag. So automating login requires Google's own supported mechanism for
+this: their public **test key pair**, which always verifies instantly
+with no real challenge, and is documented for exactly this purpose.
+
+## Required environment setup
+
+1. **Frontend build config** — set the reCAPTCHA site key to Google's
+   test key for whatever environment you run these tests against:
+   ```
+   VITE_RECAPTCHA_SITE_KEY=6LeIxAcTAAAAAJcZVRqyHh71UMIEGNQ_MXjiZKhI
+   ```
+2. **Backend config** (`AxeCode_backend`) — set the matching test secret:
+   ```
+   RECAPTCHA_SECRET_KEY=6LeIxAcTAAAAAGG-vFI1TnRWxMZNFuojJ4WifJWe
+   ```
+   **Never use this key pair in production** — anyone can pass it.
+3. **A seeded test user** on that backend, then:
+   ```
+   E2E_TEST_USER_EMAIL=your-test-user@example.com
+   E2E_TEST_USER_PASSWORD=your-test-password
+   ```
+4. Optionally, point tests at a non-default frontend URL:
+   ```
+   E2E_BASE_URL=http://localhost:5173
+   ```
+
+## Running locally
+
+```bash
+npx playwright install --with-deps chromium   # first time only
+npm run test:e2e            # headless
+npm run test:e2e:ui         # interactive UI mode, easiest for debugging
+npm run test:e2e:report     # open the HTML report from the last run
+```
+
+`playwright.config.js` auto-starts `npm run dev` locally
+(`reuseExistingServer: true`, so it won't conflict with a server you
+already have running). In CI/staging, the app and backend are assumed to
+already be deployed — see the "Full-Stack E2E" section of the QA plan for
+the staging pipeline this is meant to plug into.
diff --git a/tests/e2e/auth.spec.js b/tests/e2e/auth.spec.js
new file mode 100644
index 0000000..91e9f20
--- /dev/null
+++ b/tests/e2e/auth.spec.js
@@ -0,0 +1,34 @@
+import { test, expect } from '@playwright/test';
+import { PATHS } from '../../src/presentation/routes/paths.js';
+import { loginAsTestUser, solveTestRecaptcha } from './utils/auth.js';
+
+test.describe('Authentication', () => {
+  test('logs in with valid credentials and reaches the dashboard', async ({ page }) => {
+    await loginAsTestUser(page);
+    await expect(page).toHaveURL(PATHS.DASHBOARD);
+  });
+
+  test('shows an error and re-enables the form on invalid credentials', async ({ page }) => {
+    await page.goto(PATHS.LOGIN);
+
+    await page.getByPlaceholder('Enter your email or username').fill('nonexistent-user@example.com');
+    await page.getByPlaceholder('••••••••').fill('definitely-wrong-password');
+    await solveTestRecaptcha(page);
+
+    const submit = page.getByRole('button', { name: /Initiate Protocol/i });
+    await submit.click();
+
+    // Error banner rendered by LoginPage when loginError is set.
+    await expect(page.getByText(/./).filter({ hasText: /invalid|incorrect|failed/i }).first())
+      .toBeVisible({ timeout: 10000 });
+
+    // Must stay on the login page — a false-positive redirect here would be
+    // a serious auth bug, not just a UX issue.
+    await expect(page).toHaveURL(new RegExp(PATHS.LOGIN));
+
+    // reCAPTCHA is reset on failure (see LoginPage.handleLogin's catch
+    // block) so the submit button must go back to disabled until it's
+    // solved again.
+    await expect(submit).toBeDisabled();
+  });
+});
diff --git a/tests/e2e/courses.spec.js b/tests/e2e/courses.spec.js
new file mode 100644
index 0000000..aff22f5
--- /dev/null
+++ b/tests/e2e/courses.spec.js
@@ -0,0 +1,35 @@
+import { test, expect } from '@playwright/test';
+import { PATHS } from '../../src/presentation/routes/paths.js';
+import { loginAsTestUser } from './utils/auth.js';
+
+test.describe('Courses catalog', () => {
+  test.beforeEach(async ({ page }) => {
+    await loginAsTestUser(page);
+  });
+
+  test('opens the courses catalog while authenticated', async ({ page }) => {
+    await page.goto(PATHS.COURSES);
+    await expect(page).toHaveURL(new RegExp(PATHS.COURSES));
+  });
+
+  test('opens a course details page from the catalog', async ({ page }) => {
+    await page.goto(PATHS.COURSES);
+
+    // CourseCard always renders a "Details" button regardless of price/
+    // access state, unlike "Buy Now" / "Continue" / "Free Join" which are
+    // conditional — so it's the one stable entry point for any seed data.
+    const detailsButton = page.getByRole('button', { name: 'Details' }).first();
+    await detailsButton.waitFor({ state: 'visible', timeout: 15000 });
+    await detailsButton.click();
+
+    await expect(page).toHaveURL(new RegExp(`${PATHS.COURSES}/.+`));
+
+    // Price/access state (Buy Now vs Free Join vs Continue) depends on
+    // seed data and the logged-in test user's entitlements, so this test
+    // only asserts the details page itself rendered — not which specific
+    // purchase CTA is showing. The full purchase → entitlement → unlocked
+    // content journey is a separate, heavier scenario (P0 in the QA plan)
+    // that needs a seeded paid course + sandboxed Paymob/Kashier credentials.
+    await expect(page.locator('body')).not.toContainText('Failed to load', { timeout: 15000 });
+  });
+});
diff --git a/tests/e2e/problems.spec.js b/tests/e2e/problems.spec.js
new file mode 100644
index 0000000..b483e37
--- /dev/null
+++ b/tests/e2e/problems.spec.js
@@ -0,0 +1,41 @@
+import { test, expect } from '@playwright/test';
+import { PATHS } from '../../src/presentation/routes/paths.js';
+import { loginAsTestUser } from './utils/auth.js';
+
+test.describe('Problems workspace', () => {
+  test.beforeEach(async ({ page }) => {
+    await loginAsTestUser(page);
+  });
+
+  test('opens the problems catalog while authenticated', async ({ page }) => {
+    await page.goto(PATHS.PROBLEMS);
+    // Protected route: a successful load (not a redirect back to /login)
+    // is itself the assertion that auth + the catalog fetch both worked.
+    await expect(page).toHaveURL(new RegExp(PATHS.PROBLEMS));
+  });
+
+  test('opens a problem workspace with editor, run and submit controls', async ({ page }) => {
+    await page.goto(PATHS.PROBLEMS);
+
+    // Click the first problem card/link. The catalog UI isn't covered here
+    // (that's a separate scenario) — this test only needs *a* valid problem
+    // id, so it follows whatever the first real link on the page is rather
+    // than hardcoding one that may not exist in every environment's seed data.
+    const firstProblemLink = page.locator('a[href*="/problems/"]').first();
+    await firstProblemLink.waitFor({ state: 'visible', timeout: 15000 });
+    await firstProblemLink.click();
+
+    await expect(page).toHaveURL(/\/problems\/.+/);
+
+    // Core workspace chrome from ProblemWorkspaceHeader + ProblemPreviewPage.
+    await expect(page.getByRole('button', { name: /^Run$/ })).toBeVisible({ timeout: 15000 });
+    await expect(page.getByRole('button', { name: /^Submit$/ })).toBeVisible();
+    await expect(page.getByRole('button', { name: 'Description' })).toBeVisible();
+    await expect(page.getByRole('button', { name: 'Test Cases' })).toBeVisible();
+
+    // Monaco mounts into a `.monaco-editor` container — just verify the
+    // editor actually mounted, not its contents (that needs seeded code
+    // templates and is covered once staging test data exists).
+    await expect(page.locator('.monaco-editor')).toBeVisible({ timeout: 15000 });
+  });
+});
diff --git a/tests/e2e/utils/auth.js b/tests/e2e/utils/auth.js
new file mode 100644
index 0000000..fa17cde
--- /dev/null
+++ b/tests/e2e/utils/auth.js
@@ -0,0 +1,45 @@
+import { PATHS } from '../../../src/presentation/routes/paths.js';
+
+/**
+ * Solves Google reCAPTCHA when the app is configured with Google's public
+ * TEST site key (6LeIxAcTAAAAAJcZVRqyHh71UMIEGNQ_MXjiZKhI). With the test
+ * key, the checkbox always verifies instantly with no real challenge — it
+ * exists specifically for automated testing. See tests/e2e/README.md.
+ *
+ * No-ops (with a clear error) if the recaptcha iframe never appears, so a
+ * misconfigured environment fails loudly instead of hanging.
+ */
+export async function solveTestRecaptcha(page) {
+  const frame = page.frameLocator('iframe[title="reCAPTCHA"]').first();
+  const checkbox = frame.locator('#recaptcha-anchor');
+  await checkbox.waitFor({ state: 'visible', timeout: 15000 });
+  await checkbox.click();
+  // The test key auto-verifies almost instantly; wait for the checked state
+  // rather than a fixed timeout.
+  await frame.locator('#recaptcha-anchor[aria-checked="true"]').waitFor({ timeout: 10000 });
+}
+
+/**
+ * Logs in through the real UI (identifier + password + reCAPTCHA) and
+ * waits for the redirect to the dashboard. Requires E2E_TEST_USER_EMAIL /
+ * E2E_TEST_USER_PASSWORD to point at a seeded test account.
+ */
+export async function loginAsTestUser(page) {
+  const email = process.env.E2E_TEST_USER_EMAIL;
+  const password = process.env.E2E_TEST_USER_PASSWORD;
+
+  if (!email || !password) {
+    throw new Error(
+      'E2E_TEST_USER_EMAIL / E2E_TEST_USER_PASSWORD are not set. ' +
+      'See tests/e2e/README.md for how to seed a test account.'
+    );
+  }
+
+  await page.goto(PATHS.LOGIN);
+  await page.getByPlaceholder('Enter your email or username').fill(email);
+  await page.getByPlaceholder('••••••••').fill(password);
+  await solveTestRecaptcha(page);
+
+  await page.getByRole('button', { name: /Initiate Protocol/i }).click();
+  await page.waitForURL(PATHS.DASHBOARD, { timeout: 15000 });
+}
diff --git a/vite.config.js b/vite.config.js
index 4f89ff0..1d3278b 100644
--- a/vite.config.js
+++ b/vite.config.js
@@ -60,5 +60,18 @@ export default defineConfig(({ mode }) => {
     globals: true,
     environment: "jsdom",
     setupFiles: "./src/setupTests.js",
+    // Vitest's default include glob matches ANY *.spec.js in the repo,
+    // which would also pick up tests/e2e/** (Playwright specs, written
+    // against a different `test`/`expect` API and meant to run only via
+    // `npx playwright test`). Exclude it explicitly, alongside Vitest's
+    // normal defaults which specifying `exclude` otherwise overrides.
+    exclude: [
+      "**/node_modules/**",
+      "**/dist/**",
+      "**/cypress/**",
+      "**/.{idea,git,cache,output,temp}/**",
+      "**/{karma,rollup,webpack,vite,vitest,jest,ava,babel,nyc,cypress,tsup,build}.config.*",
+      "tests/e2e/**",
+    ],
   },
-}});
\ No newline at end of file
+}});