### Example Usage of Sandbox Provider

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Demonstrates how to use the SandboxProvider interface with the E2B provider to create a sandbox, write files, run commands, install packages, and restart the Vite server. Ensure proper cleanup by terminating the sandbox.

```typescript
// Example usage with E2B provider
const provider = SandboxFactory.create('e2b');
const info = await provider.createSandbox();

await provider.writeFile('src/App.jsx', "
  function App() {
    return <h1>Hello World</h1>;
  }
  export default App;
");

const result = await provider.runCommand('npm run build');
console.log(result.stdout, result.exitCode);

await provider.installPackages(['axios', 'lodash']);
await provider.restartViteServer();

// Cleanup
await provider.terminate();
```

--------------------------------

### Install Packages API

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Installs npm packages in the active sandbox with real-time progress streaming.

```APIDOC
## POST /api/install-packages

### Description
Installs npm packages in the active sandbox with real-time progress streaming.

### Method
POST

### Endpoint
/api/install-packages

### Parameters
#### Request Body
- **packages** (array[string]) - Required - An array of package names to install.

### Request Example
```json
{
  "packages": ["framer-motion", "@heroicons/react", "recharts"]
}
```

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the operation was successful.
- **message** (string) - A message describing the outcome of the operation.
- **installedPackages** (array[string]) - A list of packages that were successfully installed.

#### Response Example (SSE Stream)
```
data: {"type":"start","message":"Installing 3 packages...","packages":["framer-motion","@heroicons/react","recharts"]}
data: {"type":"status","message":"Stopping development server..."}
data: {"type":"status","message":"Checking installed packages..."}
data: {"type":"info","message":"Installing 3 new package(s): framer-motion, @heroicons/react, recharts"}
data: {"type":"success","message":"Successfully installed: framer-motion, @heroicons/react, recharts","installedPackages":["framer-motion","@heroicons/react","recharts"]}
data: {"type":"status","message":"Restarting development server..."}
data: {"type":"complete","message":"Package installation complete and dev server restarted!","installedPackages":["framer-motion","@heroicons/react","recharts"]}
```
```

--------------------------------

### Clone and Install Dependencies

Source: https://github.com/firecrawl/open-lovable/blob/main/README.md

Clone the Open Lovable repository and install its dependencies using pnpm. Alternatively, npm or yarn can be used.

```bash
git clone https://github.com/firecrawl/open-lovable.git
cd open-lovable
pnpm install  # or npm install / yarn install
```

--------------------------------

### Run Development Server

Source: https://github.com/firecrawl/open-lovable/blob/main/README.md

Start the Open Lovable development server using pnpm. The application will be accessible at http://localhost:3000.

```bash
pnpm dev  # or npm run dev / yarn dev
```

--------------------------------

### Install npm Packages with SSE

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Use this endpoint to install npm packages in the active sandbox. It streams real-time progress updates via Server-Sent Events (SSE).

```bash
curl -X POST http://localhost:3000/api/install-packages \
  -H "Content-Type: application/json" \
  -d '{
    "packages": ["framer-motion", "@heroicons/react", "recharts"]
  }'
```

```text
data: {"type":"start","message":"Installing 3 packages...","packages":["framer-motion","@heroicons/react","recharts"]}
data: {"type":"status","message":"Stopping development server..."}
data: {"type":"status","message":"Checking installed packages..."}
data: {"type":"info","message":"Installing 3 new package(s): framer-motion, @heroicons/react, recharts"}
data: {"type":"success","message":"Successfully installed: framer-motion, @heroicons/react, recharts","installedPackages":["framer-motion","@heroicons/react","recharts"]}
data: {"type":"status","message":"Restarting development server..."}
data: {"type":"complete","message":"Package installation complete and dev server restarted!","installedPackages":["framer-motion","@heroicons/react","recharts"]}
```

--------------------------------

### Get All Sandbox Files

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Retrieves all files from the active sandbox, including parsed component information and a file manifest. This is useful for analyzing the sandbox's contents.

```bash
curl -X GET http://localhost:3000/api/get-sandbox-files
```

```json
{
  "success": true,
  "files": {
    "src/App.jsx": "import Hero from './components/Hero';\nimport Header from './components/Header';\n...",
    "src/components/Hero.jsx": "function Hero() { return <div>...</div> }",
    "src/components/Header.jsx": "function Header() { return <header>...</header> }",
    "src/index.css": "@tailwind base;\n@tailwind components;\n@tailwind utilities;"
  },
  "structure": ".\n./src\n./src/components\n./public",
  "fileCount": 4,
  "manifest": {
    "files": {
      "/src/App.jsx": {
        "content": "...",
        "type": "component",
        "path": "/src/App.jsx",
        "relativePath": "src/App.jsx",
        "imports": ["./components/Hero", "./components/Header"],
        "exports": ["default"]
      }
    },
    "routes": [],
    "componentTree": {
      "App": ["Hero", "Header"]
    },
    "entryPoint": "/src/main.jsx",
    "styleFiles": ["/src/index.css"],
    "timestamp": 1704067200000
  }
}
```

--------------------------------

### Get Sandbox Files API

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Retrieves all files from the active sandbox with parsed component information and file manifest.

```APIDOC
## GET /api/get-sandbox-files

### Description
Retrieves all files from the active sandbox, including parsed component information and a file manifest.

### Method
GET

### Endpoint
/api/get-sandbox-files

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the files were retrieved successfully.
- **files** (object) - An object where keys are file paths and values are file contents.
- **structure** (string) - A string representing the directory structure of the sandbox.
- **fileCount** (integer) - The total number of files in the sandbox.
- **manifest** (object) - An object containing parsed information about the files, routes, component tree, entry point, and style files.

#### Response Example
```json
{
  "success": true,
  "files": {
    "src/App.jsx": "import Hero from './components/Hero';\nimport Header from './components/Header';\n...",
    "src/components/Hero.jsx": "function Hero() { return <div>...</div> }",
    "src/components/Header.jsx": "function Header() { return <header>...</header> }",
    "src/index.css": "@tailwind base;\n@tailwind components;\n@tailwind utilities;"
  },
  "structure": ".\n./src\n./src/components\n./public",
  "fileCount": 4,
  "manifest": {
    "files": {
      "/src/App.jsx": {
        "content": "...",
        "type": "component",
        "path": "/src/App.jsx",
        "relativePath": "src/App.jsx",
        "imports": ["./components/Hero", "./components/Header"],
        "exports": ["default"]
      }
    },
    "routes": [],
    "componentTree": {
      "App": ["Hero", "Header"]
    },
    "entryPoint": "/src/main.jsx",
    "styleFiles": ["/src/index.css"],
    "timestamp": 1704067200000
  }
}
```
```

--------------------------------

### Get Available Sandbox Providers

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Retrieves a list of all available sandbox providers that can be used with the SandboxFactory. This helps in determining which providers are supported.

```typescript
// Check available providers
const providers = SandboxFactory.getAvailableProviders();
// Returns: ['e2b', 'vercel']
```

--------------------------------

### Apply AI Code to Sandbox

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Applies AI-generated code to the active sandbox, handling file creation/updates, package installation, and Vite server restarts. Specify the response containing code and any new packages.

```bash
curl -X POST http://localhost:3000/api/apply-ai-code \
  -H "Content-Type: application/json" \
  -d '{
    "response": "<file path=\"src/components/Hero.jsx\">function Hero() { return <div className=\"min-h-screen bg-blue-500\"><h1>Welcome</h1></div> }</file><file path=\"src/App.jsx\">import Hero from \"./components/Hero\"; function App() { return <Hero /> } export default App;</file>",
    "isEdit": false,
    "packages": ["framer-motion"]
  }'
```

```json
{
  "success": true,
  "results": {
    "filesCreated": ["src/components/Hero.jsx", "src/App.jsx"],
    "filesUpdated": [],
    "packagesInstalled": ["framer-motion"],
    "packagesAlreadyInstalled": [],
    "packagesFailed": [],
    "commandsExecuted": [],
    "errors": []
  },
  "explanation": "Code generated successfully!",
  "message": "Applied 2 files successfully"
}
```

--------------------------------

### AI Code Application API

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Applies AI-generated code to the active sandbox, handling file creation/updates, package installation, and server restarts.

```APIDOC
## POST /api/apply-ai-code

### Description
Parses AI-generated code and writes files to the active sandbox. Handles package detection, installation, and automatic Vite server restart.

### Method
POST

### Endpoint
/api/apply-ai-code

### Parameters
#### Request Body
- **response** (string) - Required - The AI-generated code, typically in a format with file path indicators.
- **isEdit** (boolean) - Required - A flag indicating if this is an edit operation.
- **packages** (array of strings) - Optional - A list of npm packages to install.

### Request Example
```json
{
  "response": "<file path=\"src/components/Hero.jsx\">function Hero() { return <div className=\"min-h-screen bg-blue-500\"><h1>Welcome</h1></div> }</file><file path=\"src/App.jsx\">import Hero from \"./components/Hero\"; function App() { return <Hero /> } export default App;</file>",
  "isEdit": false,
  "packages": ["framer-motion"]
}
```

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the code application was successful.
- **results** (object) - Details about the applied code.
  - **filesCreated** (array of strings) - List of files created.
  - **filesUpdated** (array of strings) - List of files updated.
  - **packagesInstalled** (array of strings) - List of packages successfully installed.
  - **packagesAlreadyInstalled** (array of strings) - List of packages that were already installed.
  - **packagesFailed** (array of strings) - List of packages that failed to install.
  - **commandsExecuted** (array of strings) - List of commands executed.
  - **errors** (array of strings) - List of any errors encountered.
- **explanation** (string) - An explanation of the operation.
- **message** (string) - A summary message.

#### Response Example
```json
{
  "success": true,
  "results": {
    "filesCreated": ["src/components/Hero.jsx", "src/App.jsx"],
    "filesUpdated": [],
    "packagesInstalled": ["framer-motion"],
    "packagesAlreadyInstalled": [],
    "packagesFailed": [],
    "commandsExecuted": [],
    "errors": []
  },
  "explanation": "Code generated successfully!",
  "message": "Applied 2 files successfully"
}
```
```

--------------------------------

### Create AI Sandbox (v1 - Vercel)

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Creates a new isolated sandbox environment specifically for Vercel. This endpoint is for Vercel-only sandbox creation.

```bash
curl -X POST http://localhost:3000/api/create-ai-sandbox
```

```json
{
  "success": true,
  "sandboxId": "vercel-xyz789",
  "url": "https://xyz789.vercel-sandbox.dev:3000",
  "message": "Vercel sandbox created and Vite React app initialized"
}
```

--------------------------------

### Create AI Sandbox (v2)

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Creates a new isolated sandbox environment with a pre-configured Vite + React + Tailwind CSS application. The sandbox provider is determined by the SANDBOX_PROVIDER environment variable.

```bash
curl -X POST http://localhost:3000/api/create-ai-sandbox-v2
```

```json
{
  "success": true,
  "sandboxId": "sandbox-abc123",
  "url": "https://abc123.e2b.dev:5173",
  "provider": "e2b",
  "message": "Sandbox created and Vite React app initialized"
}
```

--------------------------------

### Create Sandbox with Environment-Configured Provider

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Instantiates a sandbox using the SandboxFactory. The factory automatically selects the provider based on environment variables. The `createSandbox` method returns information about the newly created sandbox.

```typescript
import { SandboxFactory } from '@/lib/sandbox/factory';

// Create sandbox using environment-configured provider
const provider = SandboxFactory.create();
const sandboxInfo = await provider.createSandbox();
// Returns: { sandboxId, url, provider, createdAt }
```

--------------------------------

### Create Sandbox with Explicit Provider

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Demonstrates how to explicitly specify the sandbox provider (e.g., 'e2b' or 'vercel') when creating a sandbox using the SandboxFactory.

```typescript
// Explicitly specify provider
const e2bProvider = SandboxFactory.create('e2b');
const vercelProvider = SandboxFactory.create('vercel');
```

--------------------------------

### Run Command in Sandbox (v2)

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Executes shell commands in the active sandbox environment using the v2 endpoint, which supports provider abstraction. Ensure the command is a string.

```bash
curl -X POST http://localhost:3000/api/run-command-v2 \
  -H "Content-Type: application/json" \
  -d '{
    "command": "ls -la src/components"
  }'
```

```json
{
  "success": true,
  "output": "total 24\ndrwxr-xr-x 2 user user 4096 Jan 1 12:00 .\n-rw-r--r-- 1 user user 1234 Jan 1 12:00 Hero.jsx\n-rw-r--r-- 1 user user 567 Jan 1 12:00 Header.jsx",
  "error": "",
  "exitCode": 0,
  "message": "Command executed successfully"
}
```

--------------------------------

### Run Command API

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Executes shell commands in the active sandbox environment.

```APIDOC
## POST /api/run-command-v2

### Description
Executes shell commands in the active sandbox environment using the v2 endpoint with provider abstraction.

### Method
POST

### Endpoint
/api/run-command-v2

### Parameters
#### Request Body
- **command** (string) - Required - The shell command to execute.

### Request Example
```json
{
  "command": "ls -la src/components"
}
```

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the command executed successfully.
- **output** (string) - The standard output of the command.
- **error** (string) - Any error messages produced by the command.
- **exitCode** (integer) - The exit code of the command.
- **message** (string) - A message describing the outcome of the command execution.

#### Response Example
```json
{
  "success": true,
  "output": "total 24\ndrwxr-xr-x 2 user user 4096 Jan 1 12:00 .\n-rw-r--r-- 1 user user 1234 Jan 1 12:00 Hero.jsx\n-rw-r--r-- 1 user user 567 Jan 1 12:00 Header.jsx",
  "error": "",
  "exitCode": 0,
  "message": "Command executed successfully"
}
```

## POST /api/run-command

### Description
Executes shell commands in the active sandbox environment (v1 endpoint - Vercel sandbox only).

### Method
POST

### Endpoint
/api/run-command

### Parameters
#### Request Body
- **command** (string) - Required - The shell command to execute.

### Request Example
```json
{
  "command": "npm run build"
}
```
```

--------------------------------

### Application Configuration Access

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Shows how to access and use application configuration settings for sandbox timeouts, AI models, and other behaviors. Use `getConfig` for type-safe access to configuration sections and `getConfigValue` for specific nested values.

```typescript
import { appConfig, getConfig, getConfigValue } from '@/config/app.config';

// Access configuration sections
const vercelConfig = appConfig.vercelSandbox;
// { timeoutMinutes: 15, devPort: 3000, devServerStartupDelay: 7000, ... }

const e2bConfig = appConfig.e2b;
// { timeoutMinutes: 30, vitePort: 5173, viteStartupDelay: 10000, ... }

const aiConfig = appConfig.ai;
// { defaultModel: 'google/gemini-3-pro-preview', maxTokens: 8000, ... }

// Available AI models
const models = appConfig.ai.availableModels;
// ['openai/gpt-5', 'moonshotai/kimi-k2-instruct-0905', 'anthropic/claude-sonnet-4-20250514', 'google/gemini-3-pro-preview']

// Type-safe config getter
const uiConfig = getConfig('ui');
// { showModelSelector: true, toastDuration: 3000, maxChatMessages: 100, ... }

// Nested value access
const timeout = getConfigValue('vercelSandbox.timeoutMs');
// 900000 (15 minutes in milliseconds)
```

--------------------------------

### Environment Configuration for Open Lovable

Source: https://github.com/firecrawl/open-lovable/blob/main/README.md

Configure your environment variables in a .env.local file. This includes API keys for Firecrawl, AI providers (Gemini, Anthropic, OpenAI, Groq), an optional key for Morph, and settings for the sandbox provider (Vercel or E2B).

```env
# =================================================================
# REQUIRED
# =================================================================
FIRECRAWL_API_KEY=your_firecrawl_api_key    # https://firecrawl.dev

# =================================================================
# AI PROVIDER - Choose your LLM
# =================================================================
GEMINI_API_KEY=your_gemini_api_key        # https://aistudio.google.com/app/apikey
ANTHROPIC_API_KEY=your_anthropic_api_key  # https://console.anthropic.com
OPENAI_API_KEY=your_openai_api_key        # https://platform.openai.com
GROQ_API_KEY=your_groq_api_key            # https://console.groq.com

# =================================================================
# FAST APPLY (Optional - for faster edits)
# =================================================================
MORPH_API_KEY=your_morphllm_api_key    # https://morphllm.com/dashboard

# =================================================================
# SANDBOX PROVIDER - Choose ONE: Vercel (default) or E2B
# =================================================================
SANDBOX_PROVIDER=vercel  # or 'e2b'

# Option 1: Vercel Sandbox (default)
# Choose one authentication method:

# Method A: OIDC Token (recommended for development)
# Run `vercel link` then `vercel env pull` to get VERCEL_OIDC_TOKEN automatically
VERCEL_OIDC_TOKEN=auto_generated_by_vercel_env_pull

# Method B: Personal Access Token (for production or when OIDC unavailable)
# VERCEL_TEAM_ID=team_xxxxxxxxx      # Your Vercel team ID 
# VERCEL_PROJECT_ID=prj_xxxxxxxxx    # Your Vercel project ID
# VERCEL_TOKEN=vercel_xxxxxxxxxxxx   # Personal access token from Vercel dashboard

# Option 2: E2B Sandbox
# E2B_API_KEY=your_e2b_api_key      # https://e2b.dev
```

--------------------------------

### Run Command in Sandbox (v1 - Vercel only)

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Executes shell commands in the active sandbox environment using the v1 endpoint. This endpoint is specific to the Vercel sandbox.

```bash
curl -X POST http://localhost:3000/api/run-command \
  -H "Content-Type: application/json" \
  -d '{
    "command": "npm run build"
  }'
```

--------------------------------

### Sandbox Creation API

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Endpoints for creating isolated sandbox environments for React application development.

```APIDOC
## POST /api/create-ai-sandbox-v2

### Description
Creates a new isolated sandbox environment with a pre-configured Vite + React + Tailwind CSS application. The sandbox provider is determined by the `SANDBOX_PROVIDER` environment variable (defaults to E2B).

### Method
POST

### Endpoint
/api/create-ai-sandbox-v2

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the sandbox creation was successful.
- **sandboxId** (string) - The unique identifier for the created sandbox.
- **url** (string) - The URL to access the sandbox environment.
- **provider** (string) - The sandbox provider used (e.g., 'e2b').
- **message** (string) - A confirmation message.

#### Response Example
```json
{
  "success": true,
  "sandboxId": "sandbox-abc123",
  "url": "https://abc123.e2b.dev:5173",
  "provider": "e2b",
  "message": "Sandbox created and Vite React app initialized"
}
```

## POST /api/create-ai-sandbox

### Description
Creates a new isolated sandbox environment specifically for Vercel. This is a legacy endpoint.

### Method
POST

### Endpoint
/api/create-ai-sandbox

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the sandbox creation was successful.
- **sandboxId** (string) - The unique identifier for the created sandbox.
- **url** (string) - The URL to access the sandbox environment.
- **message** (string) - A confirmation message.

#### Response Example
```json
{
  "success": true,
  "sandboxId": "vercel-xyz789",
  "url": "https://xyz789.vercel-sandbox.dev:3000",
  "message": "Vercel sandbox created and Vite React app initialized"
}
```
```

--------------------------------

### Sandbox Provider Interface Definition

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Defines the abstract interface for sandbox providers, outlining methods for file operations, command execution, and package management. Use this interface to implement custom sandbox providers.

```typescript
import { SandboxProvider, SandboxInfo, CommandResult } from '@/lib/sandbox/types';

// Provider interface methods
abstract class SandboxProvider {
  // Create and initialize sandbox
  abstract createSandbox(): Promise<SandboxInfo>;

  // Execute shell command
  abstract runCommand(command: string): Promise<CommandResult>;

  // File operations
  abstract writeFile(path: string, content: string): Promise<void>;
  abstract readFile(path: string): Promise<string>;
  abstract listFiles(directory?: string): Promise<string[]>;

  // Package management
  abstract installPackages(packages: string[]): Promise<CommandResult>;

  // Sandbox lifecycle
  abstract getSandboxUrl(): string | null;
  abstract getSandboxInfo(): SandboxInfo | null;
  abstract terminate(): Promise<void>;
  abstract isAlive(): boolean;

  // Vite-specific operations
  async setupViteApp(): Promise<void>;
  async restartViteServer(): Promise<void>;
}
```

--------------------------------

### Sandbox Provider Factory

Source: https://context7.com/firecrawl/open-lovable/llms.txt

The SandboxFactory class abstracts sandbox creation across different providers (E2B and Vercel).

```APIDOC
## SandboxFactory Class

### Description
Abstracts sandbox creation across different providers (E2B and Vercel).

### Methods
- **create(provider?: 'e2b' | 'vercel')**: Creates a sandbox instance using the specified provider or the environment-configured provider.
  - Returns: `{ sandboxId, url, provider, createdAt }`
- **getAvailableProviders()**: Returns a list of available sandbox providers.
  - Returns: `['e2b', 'vercel']`
- **isProviderAvailable(provider: 'e2b' | 'vercel')**: Checks if the specified provider has the required credentials configured.
  - Returns: `boolean`

### Usage Example (TypeScript)
```typescript
import { SandboxFactory } from '@/lib/sandbox/factory';

// Create sandbox using environment-configured provider
const provider = SandboxFactory.create();
const sandboxInfo = await provider.createSandbox();
// Returns: { sandboxId, url, provider, createdAt }

// Explicitly specify provider
const e2bProvider = SandboxFactory.create('e2b');
const vercelProvider = SandboxFactory.create('vercel');

// Check available providers
const providers = SandboxFactory.getAvailableProviders();
// Returns: ['e2b', 'vercel']

// Check if provider has required credentials
const e2bAvailable = SandboxFactory.isProviderAvailable('e2b');
// Checks for E2B_API_KEY env var

const vercelAvailable = SandboxFactory.isProviderAvailable('vercel');
// Checks for VERCEL_OIDC_TOKEN or (VERCEL_TOKEN + VERCEL_TEAM_ID + VERCEL_PROJECT_ID)
```
```

--------------------------------

### Check Provider Availability

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Checks if a specific sandbox provider is available by verifying the presence of required environment variables. For E2B, it checks for E2B_API_KEY. For Vercel, it checks for VERCEL_OIDC_TOKEN or a combination of VERCEL_TOKEN, VERCEL_TEAM_ID, and VERCEL_PROJECT_ID.

```typescript
const e2bAvailable = SandboxFactory.isProviderAvailable('e2b');
// Checks for E2B_API_KEY env var

const vercelAvailable = SandboxFactory.isProviderAvailable('vercel');
// Checks for VERCEL_OIDC_TOKEN or (VERCEL_TOKEN + VERCEL_TEAM_ID + VERCEL_PROJECT_ID)
```

--------------------------------

### Generate AI Code Stream (New App)

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Streams AI-generated React code for creating a new application based on a natural language prompt. Specify the desired model and sandbox context.

```bash
curl -X POST http://localhost:3000/api/generate-ai-code-stream \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Create a landing page with a hero section, features grid, and footer",
    "model": "anthropic/claude-sonnet-4-20250514",
    "context": {
      "sandboxId": "sandbox-abc123"
    },
    "isEdit": false
  }'
```

--------------------------------

### AI Code Generation API

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Endpoints for streaming AI-generated React code based on natural language prompts, supporting both new application generation and existing application edits.

```APIDOC
## POST /api/generate-ai-code-stream

### Description
Streams AI-generated React code based on a natural language prompt. Supports multiple AI providers and includes intelligent edit detection for modifying existing applications.

### Method
POST

### Endpoint
/api/generate-ai-code-stream

### Parameters
#### Request Body
- **prompt** (string) - Required - The natural language prompt describing the desired code.
- **model** (string) - Required - The AI model to use for generation (e.g., 'anthropic/claude-sonnet-4-20250514').
- **context** (object) - Optional - Context for the generation.
  - **sandboxId** (string) - Required if context is provided - The ID of the sandbox to use.
  - **currentFiles** (object) - Optional - A map of file paths to their current content, used for edits.
- **isEdit** (boolean) - Required - A flag indicating whether this is an edit operation (true) or a new generation (false).

### Request Example
```json
{
  "prompt": "Create a landing page with a hero section, features grid, and footer",
  "model": "anthropic/claude-sonnet-4-20250514",
  "context": {
    "sandboxId": "sandbox-abc123"
  },
  "isEdit": false
}
```

### Response
#### Success Response (200)
This endpoint uses Server-Sent Events (SSE) to stream responses.
- **type** (string) - The type of event ('status' or 'stream' or 'complete').
- **message** (string) - Status message (for 'status' type).
- **text** (string) - Generated code snippet or file path (for 'stream' type).
- **raw** (boolean) - Indicates if the text is raw code (for 'stream' type).
- **type** (string) - The type of event ('component' for component data).
- **name** (string) - The name of the component.
- **path** (string) - The file path of the component.
- **index** (number) - The index of the component.
- **type** (string) - The type of event ('complete' for completion).
- **generatedCode** (string) - The complete generated code.
- **files** (number) - The total number of files generated.
- **components** (number) - The total number of components generated.
- **model** (string) - The AI model used for generation.

#### Response Example (SSE Stream)
```
data: {"type":"status","message":"Initializing AI..."}
data: {"type":"status","message":"Planning application structure..."}
data: {"type":"stream","text":"<file path=\"src/components/Hero.jsx\">","raw":true}
data: {"type":"component","name":"Hero","path":"src/components/Hero.jsx","index":1}
data: {"type":"complete","generatedCode":"...","files":5,"components":4,"model":"anthropic/claude-sonnet-4-20250514"}
```
```

--------------------------------

### Generate AI Code Stream (Edit App)

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Streams AI-generated code to edit an existing application. Provide the prompt, model, sandbox context, and current file contents for intelligent edits.

```bash
curl -X POST http://localhost:3000/api/generate-ai-code-stream \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Change the hero background color to blue and update the headline",
    "model": "google/gemini-3-pro-preview",
    "context": {
      "sandboxId": "sandbox-abc123",
      "currentFiles": {
        "src/components/Hero.jsx": "function Hero() { return <div className=\"bg-gray-900\">...</div> }"
      }
    },
    "isEdit": true
  }'
```

--------------------------------

### Server-Sent Events Stream Response Format

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Illustrates the format of Server-Sent Events (SSE) responses during AI code generation, including status updates, code streams, and completion details.

```text
data: {"type":"status","message":"Initializing AI..."}
data: {"type":"status","message":"Planning application structure..."}
data: {"type":"stream","text":"<file path=\"src/components/Hero.jsx\">","raw":true}
data: {"type":"component","name":"Hero","path":"src/components/Hero.jsx","index":1}
data: {"type":"complete","generatedCode":"...","files":5,"components":4,"model":"anthropic/claude-sonnet-4-20250514"}
```

--------------------------------

### Scrape Website Content

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Scrapes a website using Firecrawl to extract content in specified formats. Configure options like `onlyMainContent`, `waitFor`, and `timeout` for precise scraping.

```bash
curl -X POST http://localhost:3000/api/scrape-website \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "formats": ["markdown", "html"],
    "options": {
      "onlyMainContent": true,
      "waitFor": 2000,
      "timeout": 30000
    }
  }'
```

--------------------------------

### Website Scraping API

Source: https://context7.com/firecrawl/open-lovable/llms.txt

Endpoint for scraping website content using Firecrawl, useful for cloning websites or content analysis.

```APIDOC
## POST /api/scrape-website

### Description
Scrapes a website using Firecrawl to extract content for website cloning or content analysis.

### Method
POST

### Endpoint
/api/scrape-website

### Parameters
#### Request Body
- **url** (string) - Required - The URL of the website to scrape.
- **formats** (array of strings) - Optional - The desired output formats (e.g., 'markdown', 'html'). Defaults to ['html'].
- **options** (object) - Optional - Scraping options.
  - **onlyMainContent** (boolean) - Optional - If true, only extracts the main content of the page.
  - **waitFor** (number) - Optional - Time in milliseconds to wait for the page to load.
  - **timeout** (number) - Optional - Maximum time in milliseconds to wait for the scrape operation.

### Request Example
```json
{
  "url": "https://example.com",
  "formats": ["markdown", "html"],
  "options": {
    "onlyMainContent": true,
    "waitFor": 2000,
    "timeout": 30000
  }
}
```
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.