diff --git a/README.md b/README.md
index bfc8c706..e8bc0849 100644
--- a/README.md
+++ b/README.md
@@ -241,3 +241,4 @@ Read more about E2B on the [E2B website](https://e2b.dev) and the official [E2B
 - Next.js app with LLM + Code Interpreter and streaming - [TypeScript](./examples/nextjs-code-interpreter)
 - How to run a Docker container in E2B - [Python/TypeScript](./examples/docker-in-e2b)
 - How to run Playwright in E2B - [TypeScript](./examples/playwright-in-e2b)
+- Self-healing Playwright tests with multi-model LLM fallback - [TypeScript](./examples/self-healing-playwright-tests)
diff --git a/examples/self-healing-playwright-tests/.env.template b/examples/self-healing-playwright-tests/.env.template
new file mode 100644
index 00000000..6f312bd4
--- /dev/null
+++ b/examples/self-healing-playwright-tests/.env.template
@@ -0,0 +1,21 @@
+# E2B sandbox API key — get one at https://e2b.dev
+E2B_API_KEY=
+
+# At least one model provider key is required.
+# The router will try them in the order they are configured (see ROUTER_ORDER).
+ANTHROPIC_API_KEY=
+OPENAI_API_KEY=
+GOOGLE_GENERATIVE_AI_API_KEY=
+
+# Comma-separated provider order. Default: anthropic,openai,google
+# Providers without a key are skipped automatically.
+ROUTER_ORDER=anthropic,openai,google
+
+# How many self-healing attempts before giving up. Default: 3
+HEAL_MAX_ATTEMPTS=3
+
+# Optional: pin specific models per provider.
+# Defaults are sensible for early-2026 (Opus 4.7, GPT-5, Gemini 2.5 Pro).
+ANTHROPIC_MODEL=claude-opus-4-7
+OPENAI_MODEL=gpt-5
+GOOGLE_MODEL=gemini-2.5-pro
diff --git a/examples/self-healing-playwright-tests/README.md b/examples/self-healing-playwright-tests/README.md
new file mode 100644
index 00000000..22c67ebe
--- /dev/null
+++ b/examples/self-healing-playwright-tests/README.md
@@ -0,0 +1,173 @@
+# Self-Healing Playwright Tests with Multi-Model Fallback
+
+Generate a Playwright test from a natural-language spec, run it inside an isolated E2B sandbox, and — when it fails — feed the page snapshot back to the LLM so it rewrites the test with more robust selectors. Multi-model fallback (Claude → GPT → Gemini) via the Vercel AI SDK keeps the pipeline alive if one provider is down.
+
+The pattern is inspired by the self-healing behaviour used in production at [qualitymax.io](https://qualitymax.io); this is a standalone TypeScript example contributed to the E2B Cookbook.
+
+## Why
+
+Browser tests break the moment a button moves or a class name changes. Two common responses:
+
+1. Brittle CSS selectors + a human on call — the status quo.
+2. Self-healing tests — capture the page state at failure, hand it to an LLM with the original spec, get back a fixed test.
+
+(2) only works if you can run the regenerated test in a clean, isolated environment without contaminating CI. E2B sandboxes give you exactly that: throwaway Firecracker microVMs, ~200 ms cold-start on the pre-baked `playwright-chromium` template.
+
+## Architecture
+
+```
+  TestSpec (NL description + URL)
+        │
+        ▼
+  ┌──────────────┐  fallback chain (Claude → GPT → Gemini)
+  │   router.ts  │  via Vercel AI SDK
+  └──────┬───────┘
+         │ generated Playwright TS
+         ▼
+  ┌──────────────┐  fresh sandbox per attempt
+  │   runner.ts  │  runs test in `playwright-chromium` template
+  └──────┬───────┘
+         │ pass / fail + stdout / stderr + page snapshot
+         ▼
+  ┌──────────────┐  on fail: rewrite prompt with snapshot + errors
+  │   healer.ts  │  back through router → runner, up to N attempts
+  └──────────────┘
+```
+
+Three small modules, ~250 lines total:
+
+- **`src/router.ts`** — multi-model router. Reads `ROUTER_ORDER` from env, drops providers without keys, falls back on any error.
+- **`src/runner.ts`** — E2B execution. Spins up `Sandbox.create('playwright-chromium')`, writes the generated test, installs `@playwright/test` in a user-writable work dir, executes, captures result.
+- **`src/healer.ts`** — failure → snapshot → prompt → retry loop.
+
+## Setup
+
+```bash
+npm install
+cp .env.template .env    # then fill in keys
+npm run typecheck        # sanity check
+npm run example:basic    # run the simplest example
+```
+
+You need:
+
+- An [E2B](https://e2b.dev) API key (free tier is enough).
+- At least one of: `ANTHROPIC_API_KEY`, `OPENAI_API_KEY`, `GOOGLE_GENERATIVE_AI_API_KEY`. The router silently drops providers without keys, so a partial setup still works.
+
+Per attempt wall clock: **~30–80 s** (mostly LLM latency + test run). Full healing demo with 2–3 attempts: **~2–4 min**. For CI where you want the ~5–10 s `@playwright/test` install cost gone too, build your own E2B template on top of `playwright-chromium` with it pre-baked.
+
+## Examples
+
+| Script | What it shows |
+| --- | --- |
+| `examples/01-basic-test.ts` | Simplest path: spec → generated test → one sandboxed run, no healing. |
+| `examples/02-self-healing.ts` | The headline. A vague spec produces a brittle test, the failure snapshot is fed back to the LLM, the retry passes. |
+| `examples/03-multi-model-fallback.ts` | Simulate a provider outage by poisoning the Anthropic key. The router falls back to OpenAI / Google. Needs at least two provider keys configured to demonstrate anything meaningful. |
+| `examples/04-parallel-suite.ts` | Run multiple healing tests concurrently in isolated sandboxes. |
+
+Run any of them with `npm run example:<name>` (see `package.json`).
+
+## How healing actually works
+
+The non-obvious bit is what we feed back to the LLM after a failure:
+
+1. **The runner injects a `test.afterEach` capture hook** into the generated test before each attempt (sentinel-bracketed, stripped and reattached on every heal). On any failed test it writes `await page.content()` to a known path inside the sandbox. The LLM never has to remember to do this — see "Battle-tested patterns" below for why that matters.
+2. **The runner** reads that snapshot off the sandbox filesystem after the failed run.
+3. **The healer classifies the failure** (`strict_mode_violation`, `locator_not_found`, `timeout`, `assertion_failed`, `unknown`) and steers the prompt accordingly. Strict-mode violations in particular get a dedicated hint pointing at three concrete fix patterns instead of letting the LLM re-emit equivalent fragile locators.
+4. **The heal prompt** includes the previous code, the truncated stdout/stderr, the page snapshot, and the failure-type-specific hint.
+5. **Each retry** spins up a fresh sandbox, so failures don't compound.
+
+`HEAL_MAX_ATTEMPTS` caps the loop (default `3`).
+
+### What a real run looks like
+
+```
+$ npm run example:healing
+
+→ attempt 1: LLM writes getByRole('link', { name: /Docs|Documentation/i })
+  ✘ strict mode violation: locator resolved to 4 elements
+  → healer reads /app/failure.html, feeds DOM back to the LLM
+→ attempt 2: LLM rewrites with a more specific selector
+  ✓ test passes
+
+Final: ✓ passed
+Attempts: 2
+  attempt 1 -- fail (generated by google)
+  attempt 2 -- pass (generated by google)
+```
+
+Each attempt spins up a fresh sandbox — failures from attempt 1 can't contaminate attempt 2.
+
+## Battle-tested patterns
+
+The first cut of this example trusted the LLM to do more of the wiring. Three problems showed up in production at [qualitymax.io](https://qualitymax.io) and got rolled back into the code here. If you copy this pattern into your own agent, these are the ones worth keeping:
+
+### 1. Inject the page-snapshot hook yourself; don't ask the LLM to write it
+
+Original approach: tell the LLM to call `fs.writeFileSync(somePath, await page.content())` in its own `afterEach`. Three failure modes, none loud:
+
+- LLM omits the hook in the heal pass → no snapshot → healing degrades to guessing from `stderr`.
+- LLM writes to a root-owned path (`/app/...`) → `EACCES` → silent, the test still fails for the original reason but you have no DOM.
+- LLM duplicates `import fs from 'fs'` on heal → `SyntaxError`, every retry now fails before the test runs.
+
+Fix: `src/runner.ts` injects a controlled `test.afterEach` block between sentinels (`// --- begin self-healing capture hook ---` / `// --- end ---`) and strip-and-reattaches it on every heal. The hook uses `require('fs')` inline instead of a top-level import so a duplicated injection can't break the file.
+
+### 2. Classify the failure before prompting
+
+`src/runner.ts → classifyFailure()` buckets failures into one of five types. The healer only injects the strict-mode hint when `strict_mode_violation` matched. Without classification, the strict-mode guidance pollutes prompts for unrelated failures (timeouts, assertion mismatches) and pushes the LLM toward "fixing" non-issues.
+
+Strict-mode-violation handling is the highest-leverage instance: without the explicit hint, the LLM kept swapping `.first()` in or trying a different `[class*=…]` matcher and burning the whole heal budget on the same failure. With the hint pointing at "anchor by unique text and walk up" / "scope by parent role" / "use the exact `data-test` from the snapshot", it usually heals in one extra attempt.
+
+### 3. Sentinel-based progress, not log scraping
+
+`src/runner.ts` emits `QMAX_PHASE:<phase>` markers at known points in the in-sandbox command and forwards them via the optional `onProgress` callback. A grep for a sentinel survives the npm/Playwright version bump that would break a regex over their normal log output. The pattern reads from `commands.run({ onStdout, onStderr })` rather than waiting for the run to finish, so UI consumers see progress in real time.
+
+```ts
+const result = await runInSandbox(generated, {
+  onProgress: (phase) => console.log(`[${phase}]`),
+});
+```
+
+If you want a *live browser feed* on top of progress markers (watch the test drive Chromium while it runs), see the [playwright-live-vnc-feed](../playwright-live-vnc-feed) example.
+
+## Why multi-model fallback matters
+
+LLM provider availability is not 100%. On 15 April 2026 an Anthropic incident took down Claude API access for several hours. Single-provider agents went dark; anything with a router survived.
+
+`examples/03-multi-model-fallback.ts` reproduces the failure mode locally by poisoning the Anthropic key. You need at least two provider keys configured for this to show anything meaningful — with only one key the router has nowhere to fall back to. Output on a two-key setup:
+
+```
+→ Anthropic key is poisoned. Router should fall back.
+✓ test generated -- the router landed on openai
+✓ test passed
+```
+
+Production code will want to:
+- Distinguish transient (5xx, rate limit) from permanent (auth) errors and retry transients on the same provider before falling back.
+- Track provider-level success rates and reorder dynamically.
+- Cap total wall-clock so a slow fallback doesn't block the loop.
+
+The pattern here is the floor, not the ceiling.
+
+## Notes on the `playwright-chromium` template
+
+The template ships `playwright` + browser binaries pre-installed at `/app`, but not `@playwright/test` (the test runner). The runner handles this:
+
+- `/app` is root-owned, so test files and additional npm installs go to `/home/user/work` instead.
+- `@playwright/test` is installed there, pinned to `1.51.1` to match the template's bundled Playwright version (browser/runner version skew causes "chrome-headless-shell-<rev>: cannot execute binary" failures).
+- `PLAYWRIGHT_BROWSERS_PATH` points at `/app/node_modules/playwright-core/.local-browsers` so the test runner re-uses the template's Chromium instead of downloading its own.
+
+All three bits are documented inline in `src/runner.ts`.
+
+## Adapting it
+
+- **Use a different test framework.** Replace the `ensureTestRunner` step in `src/runner.ts` and the `SYSTEM_RULES` in `src/healer.ts`. Cypress, Vitest browser mode, Pytest + Selenium — anything that runs in a Linux sandbox works.
+- **Add more providers.** `src/router.ts` maps provider names to AI SDK clients. Add a case to `modelFor()` and an entry to `ENV_KEY_BY_PROVIDER`.
+- **Build a custom E2B template.** Pre-bake `@playwright/test` on top of `playwright-chromium` to drop each attempt back toward the 200 ms cold-start floor.
+- **Wire it into CI.** The whole entry point is `runHealingTest(spec, maxAttempts)`. Call it from a GitHub Action and post the trace to your build summary.
+
+## Credits
+
+- Self-healing pattern inspired by [qualitymax.io](https://qualitymax.io).
+- Sandbox runtime: [E2B](https://e2b.dev).
+- Model orchestration: [Vercel AI SDK](https://sdk.vercel.ai).
diff --git a/examples/self-healing-playwright-tests/examples/01-basic-test.ts b/examples/self-healing-playwright-tests/examples/01-basic-test.ts
new file mode 100644
index 00000000..2df11b71
--- /dev/null
+++ b/examples/self-healing-playwright-tests/examples/01-basic-test.ts
@@ -0,0 +1,53 @@
+/**
+ * Example 1 -- the simplest path through the system.
+ *
+ * Generate a Playwright test from a one-line natural-language spec
+ * and run it once in a fresh E2B sandbox. No healing.
+ *
+ * Run:  pnpm tsx examples/01-basic-test.ts
+ */
+
+import 'dotenv/config';
+
+import { route } from '../src/router.js';
+import { runInSandbox } from '../src/runner.js';
+import type { GeneratedTest, TestSpec } from '../src/types.js';
+
+const spec: TestSpec = {
+  url: 'https://e2b.dev',
+  description:
+    'The page should mention "AI agents" somewhere on it within 10 seconds of loading.',
+};
+
+const PROMPT = `You are an expert Playwright test author.
+Output a single TypeScript file -- no markdown fences, no commentary.
+Use @playwright/test, import { test, expect }.
+
+Spec:
+URL: ${spec.url}
+What to verify: ${spec.description}
+
+Return only the TypeScript file content.`;
+
+async function main(): Promise<void> {
+  console.log('→ generating test...');
+  const { text, provider } = await route(PROMPT);
+  const generated: GeneratedTest = { code: text.trim(), provider };
+  console.log(`✓ generated by ${provider} (${generated.code.length} chars)`);
+
+  console.log('→ running in E2B playwright-chromium sandbox (~30-50s)...');
+  const result = await runInSandbox(generated);
+  console.log(result.passed ? '✓ test passed' : '✗ test failed');
+  if (!result.passed) {
+    console.log('--- stdout ---');
+    console.log(result.stdout);
+    console.log('--- stderr ---');
+    console.log(result.stderr);
+    process.exitCode = 1;
+  }
+}
+
+main().catch((err) => {
+  console.error(err);
+  process.exit(1);
+});
diff --git a/examples/self-healing-playwright-tests/examples/02-self-healing.ts b/examples/self-healing-playwright-tests/examples/02-self-healing.ts
new file mode 100644
index 00000000..85b2850e
--- /dev/null
+++ b/examples/self-healing-playwright-tests/examples/02-self-healing.ts
@@ -0,0 +1,51 @@
+/**
+ * Example 2 -- the headline example.
+ *
+ * Run a test that is *deliberately likely to fail on the first try*
+ * (description is vague enough that the LLM may pick brittle selectors).
+ * The healing loop kicks in: failure stdout + a captured page snapshot
+ * are fed back into the LLM, which rewrites the test with more robust
+ * selectors. We retry up to HEAL_MAX_ATTEMPTS times.
+ *
+ * This is the qualitymax.io self-healing pattern, ported to E2B.
+ *
+ * Run:  pnpm tsx examples/02-self-healing.ts
+ */
+
+import 'dotenv/config';
+
+import { runHealingTest } from '../src/healer.js';
+import type { TestSpec } from '../src/types.js';
+
+const spec: TestSpec = {
+  url: 'https://e2b.dev',
+  description:
+    'Click whatever link or button appears to lead to the documentation, ' +
+    'then verify the destination page contains the word "sandbox".',
+  context:
+    'The site nav may have changed since the LLM was trained. Prefer ' +
+    'role / accessible-name / data-testid based selectors.',
+};
+
+async function main(): Promise<void> {
+  const trace = await runHealingTest(spec, 3);
+
+  console.log(`\nFinal: ${trace.finalPassed ? '✓ passed' : '✗ failed'}`);
+  console.log(`Attempts: ${trace.attempts.length}`);
+
+  trace.attempts.forEach((a) => {
+    console.log(
+      `  attempt ${a.attempt} -- ${a.result.passed ? 'pass' : 'fail'} ` +
+        `(generated by ${a.generated.provider})`,
+    );
+  });
+
+  if (!trace.finalPassed) {
+    process.exitCode = 1;
+  }
+}
+
+main().catch((err) => {
+  console.error(err);
+  process.exit(1);
+});
diff --git a/examples/self-healing-playwright-tests/examples/03-multi-model-fallback.ts b/examples/self-healing-playwright-tests/examples/03-multi-model-fallback.ts
new file mode 100644
index 00000000..640bb982
--- /dev/null
+++ b/examples/self-healing-playwright-tests/examples/03-multi-model-fallback.ts
@@ -0,0 +1,52 @@
+/**
+ * Example 3 -- show the router fall back from a "broken" provider.
+ *
+ * We simulate the Anthropic outage of 15 April 2026 by overriding the
+ * Anthropic key with a guaranteed-bad value. The router should detect
+ * the auth failure and move on to OpenAI (or Google). The test still
+ * gets generated and run.
+ *
+ * This is the production behavior that kept qualitymax.io serving
+ * customers during that incident -- the architecture is the same here.
+ *
+ * Run:  pnpm tsx examples/03-multi-model-fallback.ts
+ */
+
+import 'dotenv/config';
+
+import { route } from '../src/router.js';
+import { runInSandbox } from '../src/runner.js';
+import type { GeneratedTest } from '../src/types.js';
+
+// Simulate the outage by poisoning Anthropic auth before the SDK reads it.
+// We keep OpenAI / Google keys real so the fallback chain has somewhere
+// to go.
+process.env.ANTHROPIC_API_KEY = 'sk-ant-deliberately-invalid-for-demo';
+
+const PROMPT = `You are an expert Playwright test author.
+Output a single TypeScript file -- no markdown fences, no commentary.
+Use @playwright/test, import { test, expect }.
+
+Spec:
+URL: https://e2b.dev
+What to verify: page <title> is non-empty.
+
+Return only the TypeScript file content.`;
+
+async function main(): Promise<void> {
+  console.log('→ Anthropic key is poisoned. Router should fall back.');
+  const { text, provider } = await route(PROMPT);
+  console.log(`✓ test generated -- the router landed on ${provider}`);
+
+  const generated: GeneratedTest = { code: text.trim(), provider };
+  const result = await runInSandbox(generated);
+  console.log(result.passed ? '✓ test passed' : '✗ test failed');
+  if (!result.passed) {
+    process.exitCode = 1;
+  }
+}
+
+main().catch((err) => {
+  console.error(err);
+  process.exit(1);
+});
diff --git a/examples/self-healing-playwright-tests/examples/04-parallel-suite.ts b/examples/self-healing-playwright-tests/examples/04-parallel-suite.ts
new file mode 100644
index 00000000..092fe7ad
--- /dev/null
+++ b/examples/self-healing-playwright-tests/examples/04-parallel-suite.ts
@@ -0,0 +1,50 @@
+/**
+ * Example 4 -- run multiple healing tests in parallel sandboxes.
+ *
+ * Each TestSpec gets its own isolated E2B sandbox so failures and
+ * state don't leak between them. This is roughly how a real test
+ * suite would be wired up against E2B in production.
+ *
+ * Run:  pnpm tsx examples/04-parallel-suite.ts
+ */
+
+import 'dotenv/config';
+
+import { runHealingTest } from '../src/healer.js';
+import type { TestSpec } from '../src/types.js';
+
+const suite: TestSpec[] = [
+  {
+    url: 'https://e2b.dev',
+    description: 'The page should mention "AI agents" somewhere on it.',
+  },
+  {
+    url: 'https://e2b.dev/docs',
+    description: 'There should be a link to the Cookbook somewhere on the page.',
+  },
+  {
+    url: 'https://e2b.dev/blog',
+    description: 'The page should list at least one blog post title.',
+  },
+];
+
+async function main(): Promise<void> {
+  const traces = await Promise.all(suite.map((spec) => runHealingTest(spec, 2)));
+
+  let failed = 0;
+  traces.forEach((trace) => {
+    const status = trace.finalPassed ? '✓' : '✗';
+    console.log(`${status} ${trace.spec.url} (${trace.attempts.length} attempts)`);
+    if (!trace.finalPassed) failed++;
+  });
+
+  console.log(`\n${suite.length - failed}/${suite.length} passed`);
+  if (failed > 0) {
+    process.exitCode = 1;
+  }
+}
+
+main().catch((err) => {
+  console.error(err);
+  process.exit(1);
+});
diff --git a/examples/self-healing-playwright-tests/package-lock.json b/examples/self-healing-playwright-tests/package-lock.json
new file mode 100644
index 00000000..c849badb
--- /dev/null
+++ b/examples/self-healing-playwright-tests/package-lock.json
@@ -0,0 +1,991 @@
+{
+  "name": "self-healing-playwright-tests",
+  "version": "0.1.0",
+  "lockfileVersion": 3,
+  "requires": true,
+  "packages": {
+    "": {
+      "name": "self-healing-playwright-tests",
+      "version": "0.1.0",
+      "license": "Apache-2.0",
+      "dependencies": {
+        "@ai-sdk/anthropic": "^1.2.0",
+        "@ai-sdk/google": "^1.2.0",
+        "@ai-sdk/openai": "^1.2.0",
+        "@e2b/code-interpreter": "^1.5.0",
+        "ai": "^4.3.0",
+        "dotenv": "^16.4.5",
+        "zod": "^3.23.8"
+      },
+      "devDependencies": {
+        "@types/node": "^22.0.0",
+        "tsx": "^4.19.0",
+        "typescript": "^5.6.0"
+      },
+      "engines": {
+        "node": ">=20"
+      }
+    },
+    "node_modules/@ai-sdk/anthropic": {
+      "version": "1.2.12",
+      "resolved": "https://registry.npmjs.org/@ai-sdk/anthropic/-/anthropic-1.2.12.tgz",
+      "integrity": "sha512-YSzjlko7JvuiyQFmI9RN1tNZdEiZxc+6xld/0tq/VkJaHpEzGAb1yiNxxvmYVcjvfu/PcvCxAAYXmTYQQ63IHQ==",
+      "license": "Apache-2.0",
+      "dependencies": {
+        "@ai-sdk/provider": "1.1.3",
+        "@ai-sdk/provider-utils": "2.2.8"
+      },
+      "engines": {
+        "node": ">=18"
+      },
+      "peerDependencies": {
+        "zod": "^3.0.0"
+      }
+    },
+    "node_modules/@ai-sdk/google": {
+      "version": "1.2.22",
+      "resolved": "https://registry.npmjs.org/@ai-sdk/google/-/google-1.2.22.tgz",
+      "integrity": "sha512-Ppxu3DIieF1G9pyQ5O1Z646GYR0gkC57YdBqXJ82qvCdhEhZHu0TWhmnOoeIWe2olSbuDeoOY+MfJrW8dzS3Hw==",
+      "license": "Apache-2.0",
+      "dependencies": {
+        "@ai-sdk/provider": "1.1.3",
+        "@ai-sdk/provider-utils": "2.2.8"
+      },
+      "engines": {
+        "node": ">=18"
+      },
+      "peerDependencies": {
+        "zod": "^3.0.0"
+      }
+    },
+    "node_modules/@ai-sdk/openai": {
+      "version": "1.3.24",
+      "resolved": "https://registry.npmjs.org/@ai-sdk/openai/-/openai-1.3.24.tgz",
+      "integrity": "sha512-GYXnGJTHRTZc4gJMSmFRgEQudjqd4PUN0ZjQhPwOAYH1yOAvQoG/Ikqs+HyISRbLPCrhbZnPKCNHuRU4OfpW0Q==",
+      "license": "Apache-2.0",
+      "dependencies": {
+        "@ai-sdk/provider": "1.1.3",
+        "@ai-sdk/provider-utils": "2.2.8"
+      },
+      "engines": {
+        "node": ">=18"
+      },
+      "peerDependencies": {
+        "zod": "^3.0.0"
+      }
+    },
+    "node_modules/@ai-sdk/provider": {
+      "version": "1.1.3",
+      "resolved": "https://registry.npmjs.org/@ai-sdk/provider/-/provider-1.1.3.tgz",
+      "integrity": "sha512-qZMxYJ0qqX/RfnuIaab+zp8UAeJn/ygXXAffR5I4N0n1IrvA6qBsjc8hXLmBiMV2zoXlifkacF7sEFnYnjBcqg==",
+      "license": "Apache-2.0",
+      "dependencies": {
+        "json-schema": "^0.4.0"
+      },
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@ai-sdk/provider-utils": {
+      "version": "2.2.8",
+      "resolved": "https://registry.npmjs.org/@ai-sdk/provider-utils/-/provider-utils-2.2.8.tgz",
+      "integrity": "sha512-fqhG+4sCVv8x7nFzYnFo19ryhAa3w096Kmc3hWxMQfW/TubPOmt3A6tYZhl4mUfQWWQMsuSkLrtjlWuXBVSGQA==",
+      "license": "Apache-2.0",
+      "dependencies": {
+        "@ai-sdk/provider": "1.1.3",
+        "nanoid": "^3.3.8",
+        "secure-json-parse": "^2.7.0"
+      },
+      "engines": {
+        "node": ">=18"
+      },
+      "peerDependencies": {
+        "zod": "^3.23.8"
+      }
+    },
+    "node_modules/@ai-sdk/react": {
+      "version": "1.2.12",
+      "resolved": "https://registry.npmjs.org/@ai-sdk/react/-/react-1.2.12.tgz",
+      "integrity": "sha512-jK1IZZ22evPZoQW3vlkZ7wvjYGYF+tRBKXtrcolduIkQ/m/sOAVcVeVDUDvh1T91xCnWCdUGCPZg2avZ90mv3g==",
+      "license": "Apache-2.0",
+      "dependencies": {
+        "@ai-sdk/provider-utils": "2.2.8",
+        "@ai-sdk/ui-utils": "1.2.11",
+        "swr": "^2.2.5",
+        "throttleit": "2.1.0"
+      },
+      "engines": {
+        "node": ">=18"
+      },
+      "peerDependencies": {
+        "react": "^18 || ^19 || ^19.0.0-rc",
+        "zod": "^3.23.8"
+      },
+      "peerDependenciesMeta": {
+        "zod": {
+          "optional": true
+        }
+      }
+    },
+    "node_modules/@ai-sdk/ui-utils": {
+      "version": "1.2.11",
+      "resolved": "https://registry.npmjs.org/@ai-sdk/ui-utils/-/ui-utils-1.2.11.tgz",
+      "integrity": "sha512-3zcwCc8ezzFlwp3ZD15wAPjf2Au4s3vAbKsXQVyhxODHcmu0iyPO2Eua6D/vicq/AUm/BAo60r97O6HU+EI0+w==",
+      "license": "Apache-2.0",
+      "dependencies": {
+        "@ai-sdk/provider": "1.1.3",
+        "@ai-sdk/provider-utils": "2.2.8",
+        "zod-to-json-schema": "^3.24.1"
+      },
+      "engines": {
+        "node": ">=18"
+      },
+      "peerDependencies": {
+        "zod": "^3.23.8"
+      }
+    },
+    "node_modules/@bufbuild/protobuf": {
+      "version": "2.11.0",
+      "resolved": "https://registry.npmjs.org/@bufbuild/protobuf/-/protobuf-2.11.0.tgz",
+      "integrity": "sha512-sBXGT13cpmPR5BMgHE6UEEfEaShh5Ror6rfN3yEK5si7QVrtZg8LEPQb0VVhiLRUslD2yLnXtnRzG035J/mZXQ==",
+      "license": "(Apache-2.0 AND BSD-3-Clause)"
+    },
+    "node_modules/@connectrpc/connect": {
+      "version": "2.0.0-rc.3",
+      "resolved": "https://registry.npmjs.org/@connectrpc/connect/-/connect-2.0.0-rc.3.tgz",
+      "integrity": "sha512-ARBt64yEyKbanyRETTjcjJuHr2YXorzQo0etyS5+P6oSeW8xEuzajA9g+zDnMcj1hlX2dQE93foIWQGfpru7gQ==",
+      "license": "Apache-2.0",
+      "peerDependencies": {
+        "@bufbuild/protobuf": "^2.2.0"
+      }
+    },
+    "node_modules/@connectrpc/connect-web": {
+      "version": "2.0.0-rc.3",
+      "resolved": "https://registry.npmjs.org/@connectrpc/connect-web/-/connect-web-2.0.0-rc.3.tgz",
+      "integrity": "sha512-w88P8Lsn5CCsA7MFRl2e6oLY4J/5toiNtJns/YJrlyQaWOy3RO8pDgkz+iIkG98RPMhj2thuBvsd3Cn4DKKCkw==",
+      "license": "Apache-2.0",
+      "peerDependencies": {
+        "@bufbuild/protobuf": "^2.2.0",
+        "@connectrpc/connect": "2.0.0-rc.3"
+      }
+    },
+    "node_modules/@e2b/code-interpreter": {
+      "version": "1.5.1",
+      "resolved": "https://registry.npmjs.org/@e2b/code-interpreter/-/code-interpreter-1.5.1.tgz",
+      "integrity": "sha512-mkyKjAW2KN5Yt0R1I+1lbH3lo+W/g/1+C2lnwlitXk5wqi/g94SEO41XKdmDf5WWpKG3mnxWDR5d6S/lyjmMEw==",
+      "license": "MIT",
+      "dependencies": {
+        "e2b": "^1.4.0"
+      },
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/aix-ppc64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/aix-ppc64/-/aix-ppc64-0.27.7.tgz",
+      "integrity": "sha512-EKX3Qwmhz1eMdEJokhALr0YiD0lhQNwDqkPYyPhiSwKrh7/4KRjQc04sZ8db+5DVVnZ1LmbNDI1uAMPEUBnQPg==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "aix"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/android-arm": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-arm/-/android-arm-0.27.7.tgz",
+      "integrity": "sha512-jbPXvB4Yj2yBV7HUfE2KHe4GJX51QplCN1pGbYjvsyCZbQmies29EoJbkEc+vYuU5o45AfQn37vZlyXy4YJ8RQ==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/android-arm64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-arm64/-/android-arm64-0.27.7.tgz",
+      "integrity": "sha512-62dPZHpIXzvChfvfLJow3q5dDtiNMkwiRzPylSCfriLvZeq0a1bWChrGx/BbUbPwOrsWKMn8idSllklzBy+dgQ==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/android-x64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-x64/-/android-x64-0.27.7.tgz",
+      "integrity": "sha512-x5VpMODneVDb70PYV2VQOmIUUiBtY3D3mPBG8NxVk5CogneYhkR7MmM3yR/uMdITLrC1ml/NV1rj4bMJuy9MCg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/darwin-arm64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/darwin-arm64/-/darwin-arm64-0.27.7.tgz",
+      "integrity": "sha512-5lckdqeuBPlKUwvoCXIgI2D9/ABmPq3Rdp7IfL70393YgaASt7tbju3Ac+ePVi3KDH6N2RqePfHnXkaDtY9fkw==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/darwin-x64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/darwin-x64/-/darwin-x64-0.27.7.tgz",
+      "integrity": "sha512-rYnXrKcXuT7Z+WL5K980jVFdvVKhCHhUwid+dDYQpH+qu+TefcomiMAJpIiC2EM3Rjtq0sO3StMV/+3w3MyyqQ==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/freebsd-arm64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/freebsd-arm64/-/freebsd-arm64-0.27.7.tgz",
+      "integrity": "sha512-B48PqeCsEgOtzME2GbNM2roU29AMTuOIN91dsMO30t+Ydis3z/3Ngoj5hhnsOSSwNzS+6JppqWsuhTp6E82l2w==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/freebsd-x64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/freebsd-x64/-/freebsd-x64-0.27.7.tgz",
+      "integrity": "sha512-jOBDK5XEjA4m5IJK3bpAQF9/Lelu/Z9ZcdhTRLf4cajlB+8VEhFFRjWgfy3M1O4rO2GQ/b2dLwCUGpiF/eATNQ==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/linux-arm": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-arm/-/linux-arm-0.27.7.tgz",
+      "integrity": "sha512-RkT/YXYBTSULo3+af8Ib0ykH8u2MBh57o7q/DAs3lTJlyVQkgQvlrPTnjIzzRPQyavxtPtfg0EopvDyIt0j1rA==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/linux-arm64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-arm64/-/linux-arm64-0.27.7.tgz",
+      "integrity": "sha512-RZPHBoxXuNnPQO9rvjh5jdkRmVizktkT7TCDkDmQ0W2SwHInKCAV95GRuvdSvA7w4VMwfCjUiPwDi0ZO6Nfe9A==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/linux-ia32": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-ia32/-/linux-ia32-0.27.7.tgz",
+      "integrity": "sha512-GA48aKNkyQDbd3KtkplYWT102C5sn/EZTY4XROkxONgruHPU72l+gW+FfF8tf2cFjeHaRbWpOYa/uRBz/Xq1Pg==",
+      "cpu": [
+        "ia32"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/linux-loong64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-loong64/-/linux-loong64-0.27.7.tgz",
+      "integrity": "sha512-a4POruNM2oWsD4WKvBSEKGIiWQF8fZOAsycHOt6JBpZ+JN2n2JH9WAv56SOyu9X5IqAjqSIPTaJkqN8F7XOQ5Q==",
+      "cpu": [
+        "loong64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/linux-mips64el": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-mips64el/-/linux-mips64el-0.27.7.tgz",
+      "integrity": "sha512-KabT5I6StirGfIz0FMgl1I+R1H73Gp0ofL9A3nG3i/cYFJzKHhouBV5VWK1CSgKvVaG4q1RNpCTR2LuTVB3fIw==",
+      "cpu": [
+        "mips64el"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/linux-ppc64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-ppc64/-/linux-ppc64-0.27.7.tgz",
+      "integrity": "sha512-gRsL4x6wsGHGRqhtI+ifpN/vpOFTQtnbsupUF5R5YTAg+y/lKelYR1hXbnBdzDjGbMYjVJLJTd2OFmMewAgwlQ==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/linux-riscv64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-riscv64/-/linux-riscv64-0.27.7.tgz",
+      "integrity": "sha512-hL25LbxO1QOngGzu2U5xeXtxXcW+/GvMN3ejANqXkxZ/opySAZMrc+9LY/WyjAan41unrR3YrmtTsUpwT66InQ==",
+      "cpu": [
+        "riscv64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/linux-s390x": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-s390x/-/linux-s390x-0.27.7.tgz",
+      "integrity": "sha512-2k8go8Ycu1Kb46vEelhu1vqEP+UeRVj2zY1pSuPdgvbd5ykAw82Lrro28vXUrRmzEsUV0NzCf54yARIK8r0fdw==",
+      "cpu": [
+        "s390x"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/linux-x64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-x64/-/linux-x64-0.27.7.tgz",
+      "integrity": "sha512-hzznmADPt+OmsYzw1EE33ccA+HPdIqiCRq7cQeL1Jlq2gb1+OyWBkMCrYGBJ+sxVzve2ZJEVeePbLM2iEIZSxA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/netbsd-arm64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/netbsd-arm64/-/netbsd-arm64-0.27.7.tgz",
+      "integrity": "sha512-b6pqtrQdigZBwZxAn1UpazEisvwaIDvdbMbmrly7cDTMFnw/+3lVxxCTGOrkPVnsYIosJJXAsILG9XcQS+Yu6w==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "netbsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/netbsd-x64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/netbsd-x64/-/netbsd-x64-0.27.7.tgz",
+      "integrity": "sha512-OfatkLojr6U+WN5EDYuoQhtM+1xco+/6FSzJJnuWiUw5eVcicbyK3dq5EeV/QHT1uy6GoDhGbFpprUiHUYggrw==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "netbsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/openbsd-arm64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/openbsd-arm64/-/openbsd-arm64-0.27.7.tgz",
+      "integrity": "sha512-AFuojMQTxAz75Fo8idVcqoQWEHIXFRbOc1TrVcFSgCZtQfSdc1RXgB3tjOn/krRHENUB4j00bfGjyl2mJrU37A==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openbsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/openbsd-x64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/openbsd-x64/-/openbsd-x64-0.27.7.tgz",
+      "integrity": "sha512-+A1NJmfM8WNDv5CLVQYJ5PshuRm/4cI6WMZRg1by1GwPIQPCTs1GLEUHwiiQGT5zDdyLiRM/l1G0Pv54gvtKIg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openbsd"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/openharmony-arm64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/openharmony-arm64/-/openharmony-arm64-0.27.7.tgz",
+      "integrity": "sha512-+KrvYb/C8zA9CU/g0sR6w2RBw7IGc5J2BPnc3dYc5VJxHCSF1yNMxTV5LQ7GuKteQXZtspjFbiuW5/dOj7H4Yw==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openharmony"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/sunos-x64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/sunos-x64/-/sunos-x64-0.27.7.tgz",
+      "integrity": "sha512-ikktIhFBzQNt/QDyOL580ti9+5mL/YZeUPKU2ivGtGjdTYoqz6jObj6nOMfhASpS4GU4Q/Clh1QtxWAvcYKamA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "sunos"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/win32-arm64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-arm64/-/win32-arm64-0.27.7.tgz",
+      "integrity": "sha512-7yRhbHvPqSpRUV7Q20VuDwbjW5kIMwTHpptuUzV+AA46kiPze5Z7qgt6CLCK3pWFrHeNfDd1VKgyP4O+ng17CA==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/win32-ia32": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-ia32/-/win32-ia32-0.27.7.tgz",
+      "integrity": "sha512-SmwKXe6VHIyZYbBLJrhOoCJRB/Z1tckzmgTLfFYOfpMAx63BJEaL9ExI8x7v0oAO3Zh6D/Oi1gVxEYr5oUCFhw==",
+      "cpu": [
+        "ia32"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@esbuild/win32-x64": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-x64/-/win32-x64-0.27.7.tgz",
+      "integrity": "sha512-56hiAJPhwQ1R4i+21FVF7V8kSD5zZTdHcVuRFMW0hn753vVfQN8xlx4uOPT4xoGH0Z/oVATuR82AiqSTDIpaHg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@opentelemetry/api": {
+      "version": "1.9.0",
+      "resolved": "https://registry.npmjs.org/@opentelemetry/api/-/api-1.9.0.tgz",
+      "integrity": "sha512-3giAOQvZiH5F9bMlMiv8+GSPMeqg0dbaeo58/0SlA9sxSqZhnUtxzX9/2FzyhS9sWQf5S0GJE0AKBrFqjpeYcg==",
+      "license": "Apache-2.0",
+      "engines": {
+        "node": ">=8.0.0"
+      }
+    },
+    "node_modules/@types/diff-match-patch": {
+      "version": "1.0.36",
+      "resolved": "https://registry.npmjs.org/@types/diff-match-patch/-/diff-match-patch-1.0.36.tgz",
+      "integrity": "sha512-xFdR6tkm0MWvBfO8xXCSsinYxHcqkQUlcHeSpMC2ukzOb6lwQAfDmW+Qt0AvlGd8HpsS28qKsB+oPeJn9I39jg==",
+      "license": "MIT"
+    },
+    "node_modules/@types/node": {
+      "version": "22.19.17",
+      "resolved": "https://registry.npmjs.org/@types/node/-/node-22.19.17.tgz",
+      "integrity": "sha512-wGdMcf+vPYM6jikpS/qhg6WiqSV/OhG+jeeHT/KlVqxYfD40iYJf9/AE1uQxVWFvU7MipKRkRv8NSHiCGgPr8Q==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "undici-types": "~6.21.0"
+      }
+    },
+    "node_modules/ai": {
+      "version": "4.3.19",
+      "resolved": "https://registry.npmjs.org/ai/-/ai-4.3.19.tgz",
+      "integrity": "sha512-dIE2bfNpqHN3r6IINp9znguYdhIOheKW2LDigAMrgt/upT3B8eBGPSCblENvaZGoq+hxaN9fSMzjWpbqloP+7Q==",
+      "license": "Apache-2.0",
+      "dependencies": {
+        "@ai-sdk/provider": "1.1.3",
+        "@ai-sdk/provider-utils": "2.2.8",
+        "@ai-sdk/react": "1.2.12",
+        "@ai-sdk/ui-utils": "1.2.11",
+        "@opentelemetry/api": "1.9.0",
+        "jsondiffpatch": "0.6.0"
+      },
+      "engines": {
+        "node": ">=18"
+      },
+      "peerDependencies": {
+        "react": "^18 || ^19 || ^19.0.0-rc",
+        "zod": "^3.23.8"
+      },
+      "peerDependenciesMeta": {
+        "react": {
+          "optional": true
+        }
+      }
+    },
+    "node_modules/chalk": {
+      "version": "5.6.2",
+      "resolved": "https://registry.npmjs.org/chalk/-/chalk-5.6.2.tgz",
+      "integrity": "sha512-7NzBL0rN6fMUW+f7A6Io4h40qQlG+xGmtMxfbnH/K7TAtt8JQWVQK+6g0UXKMeVJoyV5EkkNsErQ8pVD3bLHbA==",
+      "license": "MIT",
+      "engines": {
+        "node": "^12.17.0 || ^14.13 || >=16.0.0"
+      },
+      "funding": {
+        "url": "https://github.com/chalk/chalk?sponsor=1"
+      }
+    },
+    "node_modules/compare-versions": {
+      "version": "6.1.1",
+      "resolved": "https://registry.npmjs.org/compare-versions/-/compare-versions-6.1.1.tgz",
+      "integrity": "sha512-4hm4VPpIecmlg59CHXnRDnqGplJFrbLG4aFEl5vl6cK1u76ws3LLvX7ikFnTDl5vo39sjWD6AaDPYodJp/NNHg==",
+      "license": "MIT"
+    },
+    "node_modules/dequal": {
+      "version": "2.0.3",
+      "resolved": "https://registry.npmjs.org/dequal/-/dequal-2.0.3.tgz",
+      "integrity": "sha512-0je+qPKHEMohvfRTCEo3CrPG6cAzAYgmzKyxRiYSSDkS6eGJdyVJm7WaYA5ECaAD9wLB2T4EEeymA5aFVcYXCA==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=6"
+      }
+    },
+    "node_modules/diff-match-patch": {
+      "version": "1.0.5",
+      "resolved": "https://registry.npmjs.org/diff-match-patch/-/diff-match-patch-1.0.5.tgz",
+      "integrity": "sha512-IayShXAgj/QMXgB0IWmKx+rOPuGMhqm5w6jvFxmVenXKIzRqTAAsbBPT3kWQeGANj3jGgvcvv4yK6SxqYmikgw==",
+      "license": "Apache-2.0"
+    },
+    "node_modules/dotenv": {
+      "version": "16.6.1",
+      "resolved": "https://registry.npmjs.org/dotenv/-/dotenv-16.6.1.tgz",
+      "integrity": "sha512-uBq4egWHTcTt33a72vpSG0z3HnPuIl6NqYcTrKEg2azoEyl2hpW0zqlxysq2pK9HlDIHyHyakeYaYnSAwd8bow==",
+      "license": "BSD-2-Clause",
+      "engines": {
+        "node": ">=12"
+      },
+      "funding": {
+        "url": "https://dotenvx.com"
+      }
+    },
+    "node_modules/e2b": {
+      "version": "1.13.2",
+      "resolved": "https://registry.npmjs.org/e2b/-/e2b-1.13.2.tgz",
+      "integrity": "sha512-m8acE/MzMAJo1A57DakR2X1Sl5Mt1tcQO2aJfygNaQHLXby/4xsjF0UeJUB70jF7xntiR41pAMbZEHnkzrT9tw==",
+      "license": "MIT",
+      "dependencies": {
+        "@bufbuild/protobuf": "^2.6.2",
+        "@connectrpc/connect": "2.0.0-rc.3",
+        "@connectrpc/connect-web": "2.0.0-rc.3",
+        "compare-versions": "^6.1.0",
+        "openapi-fetch": "^0.9.7",
+        "platform": "^1.3.6"
+      },
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/esbuild": {
+      "version": "0.27.7",
+      "resolved": "https://registry.npmjs.org/esbuild/-/esbuild-0.27.7.tgz",
+      "integrity": "sha512-IxpibTjyVnmrIQo5aqNpCgoACA/dTKLTlhMHihVHhdkxKyPO1uBBthumT0rdHmcsk9uMonIWS0m4FljWzILh3w==",
+      "dev": true,
+      "hasInstallScript": true,
+      "license": "MIT",
+      "bin": {
+        "esbuild": "bin/esbuild"
+      },
+      "engines": {
+        "node": ">=18"
+      },
+      "optionalDependencies": {
+        "@esbuild/aix-ppc64": "0.27.7",
+        "@esbuild/android-arm": "0.27.7",
+        "@esbuild/android-arm64": "0.27.7",
+        "@esbuild/android-x64": "0.27.7",
+        "@esbuild/darwin-arm64": "0.27.7",
+        "@esbuild/darwin-x64": "0.27.7",
+        "@esbuild/freebsd-arm64": "0.27.7",
+        "@esbuild/freebsd-x64": "0.27.7",
+        "@esbuild/linux-arm": "0.27.7",
+        "@esbuild/linux-arm64": "0.27.7",
+        "@esbuild/linux-ia32": "0.27.7",
+        "@esbuild/linux-loong64": "0.27.7",
+        "@esbuild/linux-mips64el": "0.27.7",
+        "@esbuild/linux-ppc64": "0.27.7",
+        "@esbuild/linux-riscv64": "0.27.7",
+        "@esbuild/linux-s390x": "0.27.7",
+        "@esbuild/linux-x64": "0.27.7",
+        "@esbuild/netbsd-arm64": "0.27.7",
+        "@esbuild/netbsd-x64": "0.27.7",
+        "@esbuild/openbsd-arm64": "0.27.7",
+        "@esbuild/openbsd-x64": "0.27.7",
+        "@esbuild/openharmony-arm64": "0.27.7",
+        "@esbuild/sunos-x64": "0.27.7",
+        "@esbuild/win32-arm64": "0.27.7",
+        "@esbuild/win32-ia32": "0.27.7",
+        "@esbuild/win32-x64": "0.27.7"
+      }
+    },
+    "node_modules/fsevents": {
+      "version": "2.3.3",
+      "resolved": "https://registry.npmjs.org/fsevents/-/fsevents-2.3.3.tgz",
+      "integrity": "sha512-5xoDfX+fL7faATnagmWPpbFtwh/R77WmMMqqHGS65C3vvB0YHrgF+B1YmZ3441tMj5n63k0212XNoJwzlhffQw==",
+      "dev": true,
+      "hasInstallScript": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": "^8.16.0 || ^10.6.0 || >=11.0.0"
+      }
+    },
+    "node_modules/get-tsconfig": {
+      "version": "4.14.0",
+      "resolved": "https://registry.npmjs.org/get-tsconfig/-/get-tsconfig-4.14.0.tgz",
+      "integrity": "sha512-yTb+8DXzDREzgvYmh6s9vHsSVCHeC0G3PI5bEXNBHtmshPnO+S5O7qgLEOn0I5QvMy6kpZN8K1NKGyilLb93wA==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "resolve-pkg-maps": "^1.0.0"
+      },
+      "funding": {
+        "url": "https://github.com/privatenumber/get-tsconfig?sponsor=1"
+      }
+    },
+    "node_modules/json-schema": {
+      "version": "0.4.0",
+      "resolved": "https://registry.npmjs.org/json-schema/-/json-schema-0.4.0.tgz",
+      "integrity": "sha512-es94M3nTIfsEPisRafak+HDLfHXnKBhV3vU5eqPcS3flIWqcxJWgXHXiey3YrpaNsanY5ei1VoYEbOzijuq9BA==",
+      "license": "(AFL-2.1 OR BSD-3-Clause)"
+    },
+    "node_modules/jsondiffpatch": {
+      "version": "0.6.0",
+      "resolved": "https://registry.npmjs.org/jsondiffpatch/-/jsondiffpatch-0.6.0.tgz",
+      "integrity": "sha512-3QItJOXp2AP1uv7waBkao5nCvhEv+QmJAd38Ybq7wNI74Q+BBmnLn4EDKz6yI9xGAIQoUF87qHt+kc1IVxB4zQ==",
+      "license": "MIT",
+      "dependencies": {
+        "@types/diff-match-patch": "^1.0.36",
+        "chalk": "^5.3.0",
+        "diff-match-patch": "^1.0.5"
+      },
+      "bin": {
+        "jsondiffpatch": "bin/jsondiffpatch.js"
+      },
+      "engines": {
+        "node": "^18.0.0 || >=20.0.0"
+      }
+    },
+    "node_modules/nanoid": {
+      "version": "3.3.11",
+      "resolved": "https://registry.npmjs.org/nanoid/-/nanoid-3.3.11.tgz",
+      "integrity": "sha512-N8SpfPUnUp1bK+PMYW8qSWdl9U+wwNWI4QKxOYDy9JAro3WMX7p2OeVRF9v+347pnakNevPmiHhNmZ2HbFA76w==",
+      "funding": [
+        {
+          "type": "github",
+          "url": "https://github.com/sponsors/ai"
+        }
+      ],
+      "license": "MIT",
+      "bin": {
+        "nanoid": "bin/nanoid.cjs"
+      },
+      "engines": {
+        "node": "^10 || ^12 || ^13.7 || ^14 || >=15.0.1"
+      }
+    },
+    "node_modules/openapi-fetch": {
+      "version": "0.9.8",
+      "resolved": "https://registry.npmjs.org/openapi-fetch/-/openapi-fetch-0.9.8.tgz",
+      "integrity": "sha512-zM6elH0EZStD/gSiNlcPrzXcVQ/pZo3BDvC6CDwRDUt1dDzxlshpmQnpD6cZaJ39THaSmwVCxxRrPKNM1hHrDg==",
+      "license": "MIT",
+      "dependencies": {
+        "openapi-typescript-helpers": "^0.0.8"
+      }
+    },
+    "node_modules/openapi-typescript-helpers": {
+      "version": "0.0.8",
+      "resolved": "https://registry.npmjs.org/openapi-typescript-helpers/-/openapi-typescript-helpers-0.0.8.tgz",
+      "integrity": "sha512-1eNjQtbfNi5Z/kFhagDIaIRj6qqDzhjNJKz8cmMW0CVdGwT6e1GLbAfgI0d28VTJa1A8jz82jm/4dG8qNoNS8g==",
+      "license": "MIT"
+    },
+    "node_modules/platform": {
+      "version": "1.3.6",
+      "resolved": "https://registry.npmjs.org/platform/-/platform-1.3.6.tgz",
+      "integrity": "sha512-fnWVljUchTro6RiCFvCXBbNhJc2NijN7oIQxbwsyL0buWJPG85v81ehlHI9fXrJsMNgTofEoWIQeClKpgxFLrg==",
+      "license": "MIT"
+    },
+    "node_modules/react": {
+      "version": "19.2.5",
+      "resolved": "https://registry.npmjs.org/react/-/react-19.2.5.tgz",
+      "integrity": "sha512-llUJLzz1zTUBrskt2pwZgLq59AemifIftw4aB7JxOqf1HY2FDaGDxgwpAPVzHU1kdWabH7FauP4i1oEeer2WCA==",
+      "license": "MIT",
+      "peer": true,
+      "engines": {
+        "node": ">=0.10.0"
+      }
+    },
+    "node_modules/resolve-pkg-maps": {
+      "version": "1.0.0",
+      "resolved": "https://registry.npmjs.org/resolve-pkg-maps/-/resolve-pkg-maps-1.0.0.tgz",
+      "integrity": "sha512-seS2Tj26TBVOC2NIc2rOe2y2ZO7efxITtLZcGSOnHHNOQ7CkiUBfw0Iw2ck6xkIhPwLhKNLS8BO+hEpngQlqzw==",
+      "dev": true,
+      "license": "MIT",
+      "funding": {
+        "url": "https://github.com/privatenumber/resolve-pkg-maps?sponsor=1"
+      }
+    },
+    "node_modules/secure-json-parse": {
+      "version": "2.7.0",
+      "resolved": "https://registry.npmjs.org/secure-json-parse/-/secure-json-parse-2.7.0.tgz",
+      "integrity": "sha512-6aU+Rwsezw7VR8/nyvKTx8QpWH9FrcYiXXlqC4z5d5XQBDRqtbfsRjnwGyqbi3gddNtWHuEk9OANUotL26qKUw==",
+      "license": "BSD-3-Clause"
+    },
+    "node_modules/swr": {
+      "version": "2.4.1",
+      "resolved": "https://registry.npmjs.org/swr/-/swr-2.4.1.tgz",
+      "integrity": "sha512-2CC6CiKQtEwaEeNiqWTAw9PGykW8SR5zZX8MZk6TeAvEAnVS7Visz8WzphqgtQ8v2xz/4Q5K+j+SeMaKXeeQIA==",
+      "license": "MIT",
+      "dependencies": {
+        "dequal": "^2.0.3",
+        "use-sync-external-store": "^1.6.0"
+      },
+      "peerDependencies": {
+        "react": "^16.11.0 || ^17.0.0 || ^18.0.0 || ^19.0.0"
+      }
+    },
+    "node_modules/throttleit": {
+      "version": "2.1.0",
+      "resolved": "https://registry.npmjs.org/throttleit/-/throttleit-2.1.0.tgz",
+      "integrity": "sha512-nt6AMGKW1p/70DF/hGBdJB57B8Tspmbp5gfJ8ilhLnt7kkr2ye7hzD6NVG8GGErk2HWF34igrL2CXmNIkzKqKw==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=18"
+      },
+      "funding": {
+        "url": "https://github.com/sponsors/sindresorhus"
+      }
+    },
+    "node_modules/tsx": {
+      "version": "4.21.0",
+      "resolved": "https://registry.npmjs.org/tsx/-/tsx-4.21.0.tgz",
+      "integrity": "sha512-5C1sg4USs1lfG0GFb2RLXsdpXqBSEhAaA/0kPL01wxzpMqLILNxIxIOKiILz+cdg/pLnOUxFYOR5yhHU666wbw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "esbuild": "~0.27.0",
+        "get-tsconfig": "^4.7.5"
+      },
+      "bin": {
+        "tsx": "dist/cli.mjs"
+      },
+      "engines": {
+        "node": ">=18.0.0"
+      },
+      "optionalDependencies": {
+        "fsevents": "~2.3.3"
+      }
+    },
+    "node_modules/typescript": {
+      "version": "5.9.3",
+      "resolved": "https://registry.npmjs.org/typescript/-/typescript-5.9.3.tgz",
+      "integrity": "sha512-jl1vZzPDinLr9eUt3J/t7V6FgNEw9QjvBPdysz9KfQDD41fQrC2Y4vKQdiaUpFT4bXlb1RHhLpp8wtm6M5TgSw==",
+      "dev": true,
+      "license": "Apache-2.0",
+      "bin": {
+        "tsc": "bin/tsc",
+        "tsserver": "bin/tsserver"
+      },
+      "engines": {
+        "node": ">=14.17"
+      }
+    },
+    "node_modules/undici-types": {
+      "version": "6.21.0",
+      "resolved": "https://registry.npmjs.org/undici-types/-/undici-types-6.21.0.tgz",
+      "integrity": "sha512-iwDZqg0QAGrg9Rav5H4n0M64c3mkR59cJ6wQp+7C4nI0gsmExaedaYLNO44eT4AtBBwjbTiGPMlt2Md0T9H9JQ==",
+      "dev": true,
+      "license": "MIT"
+    },
+    "node_modules/use-sync-external-store": {
+      "version": "1.6.0",
+      "resolved": "https://registry.npmjs.org/use-sync-external-store/-/use-sync-external-store-1.6.0.tgz",
+      "integrity": "sha512-Pp6GSwGP/NrPIrxVFAIkOQeyw8lFenOHijQWkUTrDvrF4ALqylP2C/KCkeS9dpUM3KvYRQhna5vt7IL95+ZQ9w==",
+      "license": "MIT",
+      "peerDependencies": {
+        "react": "^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0"
+      }
+    },
+    "node_modules/zod": {
+      "version": "3.25.76",
+      "resolved": "https://registry.npmjs.org/zod/-/zod-3.25.76.tgz",
+      "integrity": "sha512-gzUt/qt81nXsFGKIFcC3YnfEAx5NkunCfnDlvuBSSFS02bcXu4Lmea0AFIUwbLWxWPx3d9p8S5QoaujKcNQxcQ==",
+      "license": "MIT",
+      "funding": {
+        "url": "https://github.com/sponsors/colinhacks"
+      }
+    },
+    "node_modules/zod-to-json-schema": {
+      "version": "3.25.2",
+      "resolved": "https://registry.npmjs.org/zod-to-json-schema/-/zod-to-json-schema-3.25.2.tgz",
+      "integrity": "sha512-O/PgfnpT1xKSDeQYSCfRI5Gy3hPf91mKVDuYLUHZJMiDFptvP41MSnWofm8dnCm0256ZNfZIM7DSzuSMAFnjHA==",
+      "license": "ISC",
+      "peerDependencies": {
+        "zod": "^3.25.28 || ^4"
+      }
+    }
+  }
+}
diff --git a/examples/self-healing-playwright-tests/package.json b/examples/self-healing-playwright-tests/package.json
new file mode 100644
index 00000000..f1a6f18a
--- /dev/null
+++ b/examples/self-healing-playwright-tests/package.json
@@ -0,0 +1,46 @@
+{
+  "name": "self-healing-playwright-tests",
+  "version": "0.1.0",
+  "description": "Cookbook example: self-healing browser tests on E2B sandboxes with Vercel AI SDK and multi-model fallback (Claude / GPT / Gemini).",
+  "license": "Apache-2.0",
+  "author": "Ruslan Strazhnyk <strazhnyk@gmail.com>",
+  "type": "module",
+  "scripts": {
+    "example:basic": "tsx examples/01-basic-test.ts",
+    "example:healing": "tsx examples/02-self-healing.ts",
+    "example:fallback": "tsx examples/03-multi-model-fallback.ts",
+    "example:suite": "tsx examples/04-parallel-suite.ts",
+    "typecheck": "tsc --noEmit"
+  },
+  "dependencies": {
+    "@ai-sdk/anthropic": "^1.2.0",
+    "@ai-sdk/google": "^1.2.0",
+    "@ai-sdk/openai": "^1.2.0",
+    "@e2b/code-interpreter": "^1.5.0",
+    "ai": "^4.3.0",
+    "dotenv": "^16.4.5",
+    "zod": "^3.23.8"
+  },
+  "devDependencies": {
+    "@types/node": "^22.0.0",
+    "tsx": "^4.19.0",
+    "typescript": "^5.6.0"
+  },
+  "engines": {
+    "node": ">=20"
+  },
+  "keywords": [
+    "e2b",
+    "ai-sdk",
+    "vercel-ai-sdk",
+    "playwright",
+    "self-healing-tests",
+    "ai-agents",
+    "multi-model",
+    "claude",
+    "gpt",
+    "gemini",
+    "cookbook",
+    "sandbox"
+  ]
+}
diff --git a/examples/self-healing-playwright-tests/src/healer.ts b/examples/self-healing-playwright-tests/src/healer.ts
new file mode 100644
index 00000000..3a5824bd
--- /dev/null
+++ b/examples/self-healing-playwright-tests/src/healer.ts
@@ -0,0 +1,173 @@
+/**
+ * Healing loop: generate test → run → on failure, ask LLM to fix it
+ * using the actual error output and a captured page snapshot, retry.
+ *
+ * Self-healing is the qualitymax.io feature this example demonstrates
+ * on E2B's substrate. The pattern is small enough to copy into any
+ * agent codebase that needs to keep brittle external tests running
+ * across UI changes.
+ */
+
+import {
+  type FailureType,
+  type GeneratedTest,
+  type HealAttempt,
+  type HealingTrace,
+  type TestSpec,
+} from './types.js';
+import { route } from './router.js';
+import { runInSandbox } from './runner.js';
+
+const DEFAULT_MAX_ATTEMPTS = Number.parseInt(
+  process.env.HEAL_MAX_ATTEMPTS ?? '3',
+  10,
+);
+
+/**
+ * Instructions the LLM gets for both first-generation and healing.
+ *
+ * Note what we *don't* ask for: the LLM no longer has to remember to
+ * dump page.content() on failure. The runner injects a controlled
+ * `test.afterEach` hook for that (see `injectCaptureHook` in runner.ts).
+ * Asking the LLM to do it produced three battle-tested failure modes
+ * at qualitymax.io — see the hook's inline comment for the full story.
+ *
+ * We also forbid markdown fences so the runner can pipe the text
+ * straight into a .ts file. Gemini in particular still occasionally
+ * wraps output in fences; the router has a defensive `stripCodeFences`
+ * pass for that, but the prompt belt-and-braces it.
+ */
+const SYSTEM_RULES = `You are an expert Playwright test author.
+Output a single TypeScript file -- no markdown fences, no commentary.
+The file must use @playwright/test and import { test, expect }.
+Prefer robust selectors: getByRole / getByText / data-testid / data-test.
+When multiple elements could match a locator, anchor by unique text and
+walk up to a container, or scope by parent role. Never use [class*=...]
+substring matchers -- they trigger strict-mode violations on real pages.`;
+
+function buildGeneratePrompt(spec: TestSpec): string {
+  return `${SYSTEM_RULES}
+
+Write a Playwright test for the following spec:
+
+URL: ${spec.url}
+What to verify: ${spec.description}
+${spec.context ? `Extra context: ${spec.context}` : ''}
+
+Return only the TypeScript file content.`;
+}
+
+/**
+ * Hint appended to the heal prompt for strict-mode violations.
+ *
+ * Battle-tested at qualitymax.io: without this, the healer treats a
+ * "resolved to 4 elements" error the same as "resolved to 0 elements"
+ * and frequently re-emits an equivalent fragile locator (e.g. swaps
+ * `.first()` in, swaps another `[class*=…]` in). Burns the whole heal
+ * budget on the same failure. Steering the LLM to one of three concrete
+ * patterns turned the strict-mode failure mode from "blocks healing"
+ * into "heals in one extra attempt" in our internal metrics.
+ */
+const STRICT_MODE_HINT = `
+=== STRICT-MODE VIOLATION DETECTED ===
+The failing locator matches multiple elements. Playwright's strict mode
+rejects ambiguous locators. Do NOT add \`.first()\` blindly -- almost always
+the right fix is one of:
+  1. Anchor by unique text and walk up to the container:
+     page.getByText('Total Balance').locator('..').getByText(/\\$\\d+/)
+  2. Scope by parent role/region:
+     page.getByRole('region', { name: 'Total Balance' }).getByText(/\\$\\d+/)
+  3. Use the exact data-test attribute if one is in the page snapshot:
+     page.locator('[data-test="total-balance"]')
+Substring class matchers like [class*="balance"] are forbidden -- they always
+hit multiple elements on a real page. Replace them with one of the above.`;
+
+function buildHealPrompt(
+  spec: TestSpec,
+  previous: GeneratedTest,
+  failure: {
+    stdout: string;
+    stderr: string;
+    snapshot?: string;
+    failureType?: FailureType;
+  },
+): string {
+  const snapshotBlock = failure.snapshot
+    ? `\nPage HTML at failure (truncated to 8000 chars):\n${failure.snapshot.slice(0, 8000)}\n`
+    : '\nNo page snapshot was captured.\n';
+
+  // Only inject the strict-mode hint when the failure is actually a
+  // strict-mode violation. Including it unconditionally pollutes prompts
+  // for unrelated failures (assertion mismatches, timeouts) and pushes
+  // the LLM toward "fixing" non-issues.
+  const failureHint = failure.failureType === 'strict_mode_violation' ? STRICT_MODE_HINT : '';
+
+  return `${SYSTEM_RULES}
+
+The previous test below failed. Fix it.
+
+Spec:
+URL: ${spec.url}
+What to verify: ${spec.description}
+${spec.context ? `Extra context: ${spec.context}` : ''}
+
+Previous test:
+${previous.code}
+
+Failure stdout:
+${failure.stdout.slice(0, 4000)}
+
+Failure stderr:
+${failure.stderr.slice(0, 2000)}
+${failureHint}
+${snapshotBlock}
+Return only the new TypeScript file content.`;
+}
+
+async function generate(spec: TestSpec): Promise<GeneratedTest> {
+  const { text, provider, usage } = await route(buildGeneratePrompt(spec));
+  return { code: text.trim(), provider, usage };
+}
+
+async function heal(
+  spec: TestSpec,
+  previous: HealAttempt,
+): Promise<GeneratedTest> {
+  const { text, provider, usage } = await route(
+    buildHealPrompt(spec, previous.generated, {
+      stdout: previous.result.stdout,
+      stderr: previous.result.stderr,
+      snapshot: previous.result.failureSnapshot,
+      failureType: previous.result.failureType,
+    }),
+  );
+  return { code: text.trim(), provider, usage };
+}
+
+/**
+ * Drive a test from natural-language spec to passing run, healing on
+ * each failure. Returns the full attempt log so the caller can inspect
+ * which provider produced which version.
+ */
+export async function runHealingTest(
+  spec: TestSpec,
+  maxAttempts: number = DEFAULT_MAX_ATTEMPTS,
+): Promise<HealingTrace> {
+  const attempts: HealAttempt[] = [];
+  let generated = await generate(spec);
+
+  for (let attempt = 1; attempt <= maxAttempts; attempt++) {
+    const result = await runInSandbox(generated);
+    attempts.push({ attempt, generated, result });
+
+    if (result.passed) {
+      return { spec, attempts, finalPassed: true };
+    }
+    if (attempt === maxAttempts) {
+      break;
+    }
+    generated = await heal(spec, attempts[attempts.length - 1]!);
+  }
+
+  return { spec, attempts, finalPassed: false };
+}
diff --git a/examples/self-healing-playwright-tests/src/index.ts b/examples/self-healing-playwright-tests/src/index.ts
new file mode 100644
index 00000000..9c07e8ab
--- /dev/null
+++ b/examples/self-healing-playwright-tests/src/index.ts
@@ -0,0 +1,21 @@
+/**
+ * Public entry points for the cookbook example.
+ */
+
+export { runHealingTest } from './healer.js';
+export { route, loadRouterConfig } from './router.js';
+export { classifyFailure, runInSandbox } from './runner.js';
+export type { RunOptions } from './runner.js';
+
+export type {
+  FailureType,
+  GeneratedTest,
+  HealAttempt,
+  HealingTrace,
+  ProviderName,
+  RouterConfig,
+  RunPhase,
+  RunResult,
+  TestSpec,
+} from './types.js';
+export { AllProvidersFailedError } from './types.js';
diff --git a/examples/self-healing-playwright-tests/src/router.ts b/examples/self-healing-playwright-tests/src/router.ts
new file mode 100644
index 00000000..cbcdbaec
--- /dev/null
+++ b/examples/self-healing-playwright-tests/src/router.ts
@@ -0,0 +1,135 @@
+/**
+ * Multi-model router with fallback chain.
+ *
+ * Tries providers in order. On rate-limit, 5xx, or network error,
+ * falls back to the next configured provider. Throws AllProvidersFailedError
+ * if every provider fails for the same call.
+ *
+ * This is the same pattern qualitymax.io uses to survive single-provider
+ * outages (the platform stayed up through the documented Anthropic
+ * outage of 15 April 2026 because the router fell back to GPT and Gemini).
+ */
+
+import { generateText } from 'ai';
+import type { LanguageModelV1 } from 'ai';
+import { anthropic } from '@ai-sdk/anthropic';
+import { openai } from '@ai-sdk/openai';
+import { google } from '@ai-sdk/google';
+
+import {
+  AllProvidersFailedError,
+  type ProviderName,
+  type RouterConfig,
+} from './types.js';
+
+const DEFAULT_MODELS: Record<ProviderName, string> = {
+  anthropic: process.env.ANTHROPIC_MODEL ?? 'claude-opus-4-7',
+  openai: process.env.OPENAI_MODEL ?? 'gpt-5',
+  google: process.env.GOOGLE_MODEL ?? 'gemini-2.5-pro',
+};
+
+const ENV_KEY_BY_PROVIDER: Record<ProviderName, string> = {
+  anthropic: 'ANTHROPIC_API_KEY',
+  openai: 'OPENAI_API_KEY',
+  google: 'GOOGLE_GENERATIVE_AI_API_KEY',
+};
+
+function modelFor(provider: ProviderName, override?: string): LanguageModelV1 {
+  const id = override ?? DEFAULT_MODELS[provider];
+  switch (provider) {
+    case 'anthropic':
+      return anthropic(id);
+    case 'openai':
+      return openai(id);
+    case 'google':
+      return google(id);
+  }
+}
+
+/**
+ * Read the router config from env vars, with sensible defaults.
+ *
+ * Providers without a configured API key are silently dropped from the
+ * order so that a partial setup still works.
+ */
+export function loadRouterConfig(): RouterConfig {
+  const raw = process.env.ROUTER_ORDER ?? 'anthropic,openai,google';
+  const requested = raw
+    .split(',')
+    .map((s) => s.trim().toLowerCase())
+    .filter((s): s is ProviderName =>
+      s === 'anthropic' || s === 'openai' || s === 'google',
+    );
+  const order = requested.filter((p) => Boolean(process.env[ENV_KEY_BY_PROVIDER[p]]));
+  if (order.length === 0) {
+    throw new Error(
+      'No usable provider keys found. Set at least one of ' +
+        `${Object.values(ENV_KEY_BY_PROVIDER).join(', ')} in your environment.`,
+    );
+  }
+  return { order };
+}
+
+export interface RouteCallResult {
+  text: string;
+  provider: ProviderName;
+  usage?: { promptTokens?: number; completionTokens?: number };
+}
+
+/**
+ * Strip a single outer markdown fence if present.
+ *
+ * LLMs occasionally wrap their response in ```lang ... ``` despite
+ * prompts that explicitly forbid it (Gemini is the usual offender).
+ * Since every caller in this cookbook expects raw TypeScript that can
+ * be written straight to a .ts file, we normalise at the router.
+ *
+ * Surgical by design: only strips when the trimmed text starts AND
+ * ends with a fence, so response bodies that legitimately contain
+ * inline fences are left alone.
+ */
+function stripCodeFences(text: string): string {
+  const trimmed = text.trim();
+  const match = trimmed.match(/^```[^\n]*\n([\s\S]*?)\n?```$/);
+  return match ? match[1]! : trimmed;
+}
+
+/**
+ * Try the configured providers in order until one returns a response.
+ *
+ * The error classification here is intentionally conservative -- on any
+ * thrown error we move on to the next provider. Production code may want
+ * to retry transient errors on the same provider before falling back.
+ */
+export async function route(
+  prompt: string,
+  config: RouterConfig = loadRouterConfig(),
+): Promise<RouteCallResult> {
+  const failures: { provider: ProviderName; error: Error }[] = [];
+
+  for (const provider of config.order) {
+    try {
+      const model = modelFor(provider, config.models?.[provider]);
+      const { text, usage } = await generateText({
+        model,
+        prompt,
+        // Slightly lower temperature to keep generated test code stable.
+        temperature: 0.2,
+      });
+      return {
+        text: stripCodeFences(text),
+        provider,
+        usage: {
+          promptTokens: usage?.promptTokens,
+          completionTokens: usage?.completionTokens,
+        },
+      };
+    } catch (err) {
+      const e = err instanceof Error ? err : new Error(String(err));
+      failures.push({ provider, error: e });
+      // Loop continues to the next provider in `order`.
+    }
+  }
+
+  throw new AllProvidersFailedError(failures);
+}
diff --git a/examples/self-healing-playwright-tests/src/runner.ts b/examples/self-healing-playwright-tests/src/runner.ts
new file mode 100644
index 00000000..9151d388
--- /dev/null
+++ b/examples/self-healing-playwright-tests/src/runner.ts
@@ -0,0 +1,297 @@
+/**
+ * E2B sandbox runner: writes a generated Playwright test file into a
+ * fresh sandbox on E2B's `playwright-chromium` template (Playwright +
+ * Chromium + system deps pre-baked), executes the test, and captures
+ * pass/fail plus stdout/stderr.
+ *
+ * Each call spins up a new sandbox so failures are isolated. Cold start
+ * is ~200ms on the pre-baked template; swap to `base` and add an
+ * `installPlaywright` step if you need to demonstrate bring-your-own.
+ */
+
+import { CommandExitError, Sandbox } from '@e2b/code-interpreter';
+
+import type { GeneratedTest, RunPhase, RunResult } from './types.js';
+
+const TEMPLATE = 'playwright-chromium';
+
+// Two-directory layout: the template's /app is owned by root and
+// contains the pre-baked Playwright browser binaries, so we can't
+// write our test files or install @playwright/test there. /home/user
+// is user-writable; we keep tests and the test-runner package here and
+// point PLAYWRIGHT_BROWSERS_PATH at /app's browser cache.
+const WORK_DIR = '/home/user/work';
+const BROWSERS_PATH = '/app/node_modules/playwright-core/.local-browsers';
+const TEST_FILE_PATH = `${WORK_DIR}/test.spec.ts`;
+const SNAPSHOT_PATH = `${WORK_DIR}/failure.html`;
+
+// Pin @playwright/test to match the version of `playwright` bundled in
+// the template (check with `cat /app/package.json` in a probe sandbox).
+// Version skew between the test runner and the browser binaries causes
+// "chrome-headless-shell-<rev>: cannot execute binary" type failures.
+const PLAYWRIGHT_VERSION = '1.51.1';
+
+// Sentinels used to bracket the injected page-snapshot hook so we can
+// strip-and-reattach it on every heal pass. Without strip-and-reattach,
+// successive heals either:
+//   (a) drop the hook entirely → no snapshot → healing degrades to
+//       guessing from stderr; or
+//   (b) duplicate the hook → "Identifier '_fs' has already been
+//       declared" SyntaxError → every retry fails before the test runs.
+// Battle-tested at qualitymax.io after both modes bit us in production.
+const HOOK_BEGIN = '// --- begin self-healing capture hook ---';
+const HOOK_END = '// --- end self-healing capture hook ---';
+
+// We inline `require('fs')` inside the callback rather than adding a
+// top-level `import fs from 'fs'`. Two reasons:
+//   1. The LLM-generated code may or may not already import fs — adding
+//      our own import risks a duplicate binding.
+//   2. If the LLM accidentally pastes the hook twice during a heal, two
+//      top-level `_fs` bindings produce a SyntaxError that's invisible
+//      until the next run. `require()` is cached and binding-free, so
+//      duplication is harmless (the strip-and-reattach still runs).
+const CAPTURE_HOOK = `${HOOK_BEGIN}
+test.afterEach(async ({ page }, testInfo) => {
+  if (testInfo.status !== testInfo.expectedStatus) {
+    try {
+      const _html = await page.content();
+      // eslint-disable-next-line @typescript-eslint/no-require-imports
+      require('fs').writeFileSync('${SNAPSHOT_PATH}', _html.substring(0, 60000));
+    } catch (_e) {
+      // Best-effort: a snapshot failure must never mask the real test failure.
+    }
+  }
+});
+${HOOK_END}`;
+
+const HOOK_STRIP_RE = new RegExp(
+  `\\n*${HOOK_BEGIN.replace(/[.*+?^${}()|[\\]\\\\]/g, '\\\\$&')}[\\s\\S]*?${HOOK_END.replace(
+    /[.*+?^${}()|[\\]\\\\]/g,
+    '\\\\$&',
+  )}\\n*`,
+  'g',
+);
+
+/**
+ * Remove a previously injected capture hook (idempotent).
+ * Always strip then reattach on each heal — see HOOK_BEGIN comment.
+ */
+function stripCaptureHook(code: string): string {
+  return code.includes(HOOK_BEGIN) ? code.replace(HOOK_STRIP_RE, '\n') : code;
+}
+
+/**
+ * Append our controlled capture hook to the LLM-generated test.
+ *
+ * We do this server-side instead of asking the LLM to write the snapshot
+ * itself (the obvious approach). Battle-tested reason: when the LLM is
+ * responsible for the snapshot path, you accumulate three failure modes
+ * none of which fail loudly:
+ *
+ *   1. LLM forgets the hook in the heal pass.
+ *   2. LLM writes to a root-owned directory (`/app/...`) → EACCES,
+ *      silent: the test fails for unrelated reasons and you have no DOM.
+ *   3. LLM duplicates `import fs from 'fs'` on heal → SyntaxError.
+ *
+ * Injecting the hook ourselves makes the snapshot a guaranteed side-effect
+ * of any failure, not a thing the agent might or might not remember to do.
+ */
+function injectCaptureHook(code: string): string {
+  return `${stripCaptureHook(code).trimEnd()}\n\n${CAPTURE_HOOK}\n`;
+}
+
+export const _internals = { stripCaptureHook, injectCaptureHook, CAPTURE_HOOK };
+
+/**
+ * Write the generated test into the sandbox along with a small Playwright
+ * config that drops the test runner into the working directory.
+ */
+async function prepareSandbox(sandbox: Sandbox, code: string): Promise<void> {
+  await sandbox.files.write(TEST_FILE_PATH, injectCaptureHook(code));
+  await sandbox.files.write(
+    `${WORK_DIR}/playwright.config.ts`,
+    `import { defineConfig } from '@playwright/test';
+export default defineConfig({
+  testDir: '${WORK_DIR}',
+  reporter: 'list',
+  use: { headless: true },
+});
+`,
+  );
+}
+
+/**
+ * The `playwright-chromium` template ships with `playwright` + browser
+ * binaries pre-installed at /app, but not `@playwright/test` (the test
+ * runner). Create a user-writable work dir and install just that --
+ * cheap (~5-10s) compared to a full browser install.
+ *
+ * If E2B later adds `@playwright/test` to the template at a location
+ * we can resolve from, this function can be deleted.
+ */
+async function ensureTestRunner(sandbox: Sandbox): Promise<void> {
+  await sandbox.commands.run(
+    `mkdir -p ${WORK_DIR} && ` +
+      `cd ${WORK_DIR} && ` +
+      `npm init -y >/dev/null && ` +
+      `npm install --no-audit --no-fund --silent @playwright/test@${PLAYWRIGHT_VERSION}`,
+    { timeoutMs: 120 * 1000 },
+  );
+}
+
+/**
+ * Best-effort read of the page snapshot the injected hook writes on failure.
+ *
+ * We control where the snapshot lives (`SNAPSHOT_PATH`) and we control the
+ * hook that writes it, so the only reasons the file is missing are:
+ *   - the test crashed before page.content() resolved (e.g. browser launch
+ *     failure); or
+ *   - the failing fixture was so broken Playwright never reached afterEach.
+ * In both cases we degrade gracefully and let the healer work from stderr.
+ */
+async function tryReadSnapshot(sandbox: Sandbox): Promise<string | undefined> {
+  try {
+    const content = await sandbox.files.read(SNAPSHOT_PATH);
+    return typeof content === 'string' && content.length > 0 ? content : undefined;
+  } catch {
+    return undefined;
+  }
+}
+
+export interface RunOptions {
+  /** Optional progress callback fired on each in-sandbox phase boundary. */
+  onProgress?: (phase: RunPhase, detail?: string) => void;
+  /** Optional callback for raw stdout chunks as the sandbox emits them. */
+  onStdoutChunk?: (chunk: string) => void;
+  /** Optional callback for raw stderr chunks as the sandbox emits them. */
+  onStderrChunk?: (chunk: string) => void;
+}
+
+/**
+ * Parse phase markers out of streamed output.
+ *
+ * The in-sandbox runner echoes `QMAX_PHASE:<phase>` lines at known
+ * boundaries; we surface those to the optional `onProgress` callback so
+ * UI consumers can show "Installing @playwright/test…", "Running test…"
+ * etc. without parsing arbitrary npm/Playwright output. The convention
+ * is borrowed from qualitymax.io's progress bar — a sentinel grep
+ * survives version bumps of npm/playwright that would break a regex on
+ * their normal log output.
+ */
+const PHASE_RE = /QMAX_PHASE:([a-z_]+)/g;
+
+function parsePhases(
+  chunk: string,
+  onProgress?: (phase: RunPhase, detail?: string) => void,
+): void {
+  if (!onProgress) return;
+  let match: RegExpExecArray | null;
+  PHASE_RE.lastIndex = 0;
+  while ((match = PHASE_RE.exec(chunk)) !== null) {
+    onProgress(match[1] as RunPhase);
+  }
+}
+
+/**
+ * Run a generated test once in a fresh sandbox.
+ *
+ * Caller is responsible for retry / healing -- this function returns
+ * structured output for one attempt only.
+ */
+export async function runInSandbox(
+  generated: GeneratedTest,
+  opts: RunOptions = {},
+): Promise<RunResult> {
+  opts.onProgress?.('sandbox_starting');
+  const sandbox = await Sandbox.create(TEMPLATE);
+  try {
+    await prepareSandbox(sandbox, generated.code);
+    opts.onProgress?.('project_uploaded');
+    await ensureTestRunner(sandbox);
+    opts.onProgress?.('deps_installed');
+
+    // E2B's SDK throws CommandExitError on non-zero exit rather than
+    // returning a result. For a test runner that's the wrong default --
+    // a failed test IS the signal we want to capture and feed into the
+    // healing loop. Catch the typed error and use it directly, since
+    // CommandExitError implements CommandResult (exitCode/stdout/stderr
+    // are getters on the instance itself).
+    //
+    // PLAYWRIGHT_BROWSERS_PATH points at the template's pre-installed
+    // Chromium so @playwright/test doesn't try to download its own.
+    //
+    // The `echo QMAX_PHASE:test_started` markers are picked up by
+    // `parsePhases` and forwarded to onProgress -- see the comment on
+    // PHASE_RE for why we use sentinels instead of parsing npm output.
+    const cmd =
+      `echo QMAX_PHASE:test_started && ` +
+      `PLAYWRIGHT_BROWSERS_PATH=${BROWSERS_PATH} ` +
+      `npx playwright test --reporter=list 2>&1; ` +
+      `status=$?; echo QMAX_PHASE:test_finished; exit $status`;
+
+    let exec: { exitCode: number; stdout: string; stderr: string };
+    try {
+      exec = await sandbox.commands.run(cmd, {
+        cwd: WORK_DIR,
+        timeoutMs: 5 * 60 * 1000,
+        onStdout: (chunk: string) => {
+          opts.onStdoutChunk?.(chunk);
+          parsePhases(chunk, opts.onProgress);
+        },
+        onStderr: (chunk: string) => {
+          opts.onStderrChunk?.(chunk);
+          parsePhases(chunk, opts.onProgress);
+        },
+      });
+    } catch (err) {
+      if (!(err instanceof CommandExitError)) throw err;
+      exec = err;
+    }
+
+    const passed = exec.exitCode === 0;
+    const failureSnapshot = passed ? undefined : await tryReadSnapshot(sandbox);
+    opts.onProgress?.('artifacts_collected');
+
+    return {
+      passed,
+      exitCode: exec.exitCode,
+      stdout: exec.stdout,
+      stderr: exec.stderr,
+      failureSnapshot,
+      failureType: passed ? undefined : classifyFailure(exec.stdout, exec.stderr),
+    };
+  } finally {
+    await sandbox.kill();
+  }
+}
+
+/**
+ * Classify a Playwright failure into one of a small set of buckets.
+ *
+ * The buckets are tuned for what the healer can actually act on:
+ *   - `strict_mode_violation` → tell the LLM how to disambiguate
+ *     (anchor by unique text, scope by parent role, use data-testid).
+ *   - `locator_not_found` → broaden the selector strategy.
+ *   - `timeout` → wait for network/load state, not bare element.
+ *   - `assertion_failed` → likely the expected value changed, not the
+ *     locator; tell the LLM to use a less specific assertion.
+ *
+ * Battle-tested ordering: strict-mode check before "locator …" because
+ * strict-mode errors *contain* "locator" in their message and would
+ * otherwise be mis-bucketed.
+ */
+export function classifyFailure(
+  stdout: string,
+  stderr: string,
+): RunResult['failureType'] {
+  const text = `${stdout}\n${stderr}`.toLowerCase();
+  if (/strict mode violation.*resolved to \d+ elements?/.test(text)) {
+    return 'strict_mode_violation';
+  }
+  if (/locator|selector|waiting for/.test(text)) {
+    return 'locator_not_found';
+  }
+  if (/timeout/.test(text)) return 'timeout';
+  if (/assertion|expect\(/.test(text)) return 'assertion_failed';
+  return 'unknown';
+}
diff --git a/examples/self-healing-playwright-tests/src/types.ts b/examples/self-healing-playwright-tests/src/types.ts
new file mode 100644
index 00000000..6d06c6ac
--- /dev/null
+++ b/examples/self-healing-playwright-tests/src/types.ts
@@ -0,0 +1,98 @@
+/**
+ * Shared types for the self-healing test runner.
+ *
+ * The flow:
+ *   1. caller describes a test in natural language       → TestSpec
+ *   2. router asks an LLM to write Playwright code       → GeneratedTest
+ *   3. runner executes the test inside an E2B sandbox    → RunResult
+ *   4. on failure, healer asks the LLM to fix it         → HealAttempt
+ *   5. loop until pass or HEAL_MAX_ATTEMPTS is hit       → HealingTrace
+ */
+
+export type ProviderName = 'anthropic' | 'openai' | 'google';
+
+/**
+ * Coarse classification of the failure so the healer can steer the next
+ * prompt. Battle-tested at qualitymax.io: without this, the healer treats
+ * "selector matches 4 elements" and "selector matches 0 elements" the
+ * same way and keeps emitting equivalent locators.
+ */
+export type FailureType =
+  | 'strict_mode_violation'
+  | 'locator_not_found'
+  | 'timeout'
+  | 'assertion_failed'
+  | 'unknown';
+
+/**
+ * Sentinel phase emitted by the in-sandbox runner script. Surfaced via
+ * the optional `onProgress` callback on `runInSandbox` so callers driving
+ * a UI / CI can show real progress instead of an opaque "running…".
+ */
+export type RunPhase =
+  | 'sandbox_starting'
+  | 'project_uploaded'
+  | 'deps_installed'
+  | 'test_started'
+  | 'test_finished'
+  | 'artifacts_collected';
+
+export interface TestSpec {
+  /** Natural-language description of what the test should do. */
+  description: string;
+  /** Target URL the test should drive. */
+  url: string;
+  /** Optional extra context for the LLM (selectors, expected text, etc.). */
+  context?: string;
+}
+
+export interface GeneratedTest {
+  /** TypeScript Playwright test file content. */
+  code: string;
+  /** Provider that produced the code. */
+  provider: ProviderName;
+  /** Token usage if reported by the provider. */
+  usage?: { promptTokens?: number; completionTokens?: number };
+}
+
+export interface RunResult {
+  passed: boolean;
+  exitCode: number;
+  stdout: string;
+  stderr: string;
+  /** Snapshot of the failing page (HTML), if the runner managed to capture one. */
+  failureSnapshot?: string;
+  /** Coarse classification of the failure, populated when !passed. */
+  failureType?: FailureType;
+}
+
+export interface HealAttempt {
+  attempt: number;
+  generated: GeneratedTest;
+  result: RunResult;
+}
+
+export interface HealingTrace {
+  spec: TestSpec;
+  attempts: HealAttempt[];
+  finalPassed: boolean;
+}
+
+export interface RouterConfig {
+  /** Order in which providers are tried. */
+  order: ProviderName[];
+  /** Per-provider model id override. */
+  models?: Partial<Record<ProviderName, string>>;
+}
+
+export class AllProvidersFailedError extends Error {
+  constructor(
+    public attempts: { provider: ProviderName; error: Error }[],
+  ) {
+    const summary = attempts
+      .map((a) => `${a.provider}: ${a.error.message}`)
+      .join(' | ');
+    super(`All configured providers failed -- ${summary}`);
+    this.name = 'AllProvidersFailedError';
+  }
+}
diff --git a/examples/self-healing-playwright-tests/tsconfig.json b/examples/self-healing-playwright-tests/tsconfig.json
new file mode 100644
index 00000000..732ebadb
--- /dev/null
+++ b/examples/self-healing-playwright-tests/tsconfig.json
@@ -0,0 +1,20 @@
+{
+  "compilerOptions": {
+    "target": "ES2022",
+    "module": "ESNext",
+    "moduleResolution": "Bundler",
+    "lib": ["ES2022"],
+    "strict": true,
+    "esModuleInterop": true,
+    "skipLibCheck": true,
+    "forceConsistentCasingInFileNames": true,
+    "resolveJsonModule": true,
+    "isolatedModules": true,
+    "noUncheckedIndexedAccess": true,
+    "noImplicitOverride": true,
+    "noFallthroughCasesInSwitch": true,
+    "allowSyntheticDefaultImports": true
+  },
+  "include": ["src/**/*", "examples/**/*"],
+  "exclude": ["node_modules", "dist"]
+}