### GET /scalar

Source: https://github.com/aytymchuk/dilcore-ai-core/blob/main/README.md

Provides access to interactive API documentation generated by Scalar.

```APIDOC
## GET /scalar

### Description
Access interactive API documentation generated by Scalar.

### Method
GET

### Endpoint
/scalar

### Response
#### Success Response (200)
Returns the interactive API documentation UI.

#### Response Example
(This endpoint serves HTML/JavaScript for the interactive documentation UI, not a JSON response.)
```

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

### Environment Variable Configuration for OpenRouter (Bash)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Provides an example of how to configure settings using environment variables, specifically for the OpenRouter service. This includes setting the API key, base URL, model, and other application-level settings like debug mode and log level.

```bash
# .env file configuration
OPENROUTER__API_KEY=sk-or-v1-your-api-key-here
OPENROUTER__BASE_URL=https://openrouter.ai/api/v1
OPENROUTER__MODEL=openai/gpt-oss-20b:free
APP_DEBUG=false
LOG_LEVEL=INFO
```

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

### GET /

Source: https://github.com/aytymchuk/dilcore-ai-core/blob/main/README.md

The root endpoint of the API, typically providing a link to the main API documentation.

```APIDOC
## GET /

### Description
Root endpoint, usually redirects to or provides a link to the main API documentation.

### Method
GET

### Endpoint
/

### Response
#### Success Response (200)
Typically returns an HTML page with a link to the API documentation (e.g., Scalar docs).

#### Response Example
(This endpoint usually serves HTML content.)
```

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

### GET /api/v1/health

Source: https://github.com/aytymchuk/dilcore-ai-core/blob/main/README.md

Performs a health check on the API service to ensure it is running and responsive.

```APIDOC
## GET /api/v1/health

### Description
Health check endpoint for the API service.

### Method
GET

### Endpoint
/api/v1/health

### Response
#### Success Response (200)
- **status** (string) - Indicates the health status of the service (e.g., "ok").

#### Response Example
```json
{
  "status": "ok"
}
```
```

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

### GET /api/v1/health

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Checks the health status of the AI Template Agent service.

```APIDOC
## GET /api/v1/health

### Description
This endpoint provides a simple health check for the AI Template Agent service. It returns a status indicating whether the service is operational.

### Method
GET

### Endpoint
`/api/v1/health`

### Response
#### Success Response (200 OK)
- **status** (string) - Indicates the health status, typically "ok".

#### Response Example
```json
{
  "status": "ok"
}
```
```

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

### Check AI Template Agent Health (Bash)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

This snippet demonstrates how to check the health of the AI Template Agent service using curl. It sends a GET request to the health check endpoint and expects a JSON response indicating the service status, application name, and the active AI model.

```bash
# Check service health
curl http://localhost:8000/api/v1/health

# Expected response
{
  "status": "healthy",
  "app_name": "AI Template Agent",
  "model": "openai/gpt-oss-20b:free"
}
```

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

### Create Agent and Generate Template (Python)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Demonstrates creating an agent using a factory function (singleton pattern) and asynchronously generating a template. It then accesses and prints properties of the generated template, including its ID, name, and sections with fields.

```python
settings = get_settings()
agent = create_template_agent(settings)

template = await agent.generate_template("Create an employee onboarding checklist")

print(f"Template ID: {template.template_id}")
print(f"Template Name: {template.template_name}")
print(f"Sections: {len(template.sections)}")

for section in template.sections:
    print(f"  Section: {section.title}")
    for field in section.fields:
        print(f"    - {field.name} ({field.type}, required={field.required})")
```

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

### Initialize TemplateAgent (Python)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

This Python snippet shows the import statements for the `TemplateAgent` class and the `create_template_agent` function, along with the `get_settings` function for configuration. These are the core components for setting up and running the synchronous template generation functionality.

```python
from ai_agent.agent import TemplateAgent, create_template_agent
from ai_agent.config import get_settings
```

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

### Configuration Settings Loading and Access (Python)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Shows how to load and access application configuration settings using Pydantic Settings. It demonstrates accessing general settings like app name and debug mode, as well as nested settings for external services like OpenRouter, including secure handling of API keys.

```python
from ai_agent.config import Settings, get_settings

settings = get_settings()

print(f"App Name: {settings.app_name}")
print(f"Debug Mode: {settings.app_debug}")
print(f"Log Level: {settings.log_level}")

print(f"Model: {settings.openrouter.model}")
print(f"Base URL: {settings.openrouter.base_url}")
api_key = settings.openrouter.api_key.get_secret_value()
```

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

### Run FastAPI Development Server and Tests

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Provides bash commands for running the FastAPI development server with Uvicorn, including options for custom ports and debug logging. It also includes commands for running tests with Pytest, with and without code coverage.

```bash
# Start development server
uv run uvicorn src.ai_agent.main:app --reload

# Start on custom port with debug logging
LOG_LEVEL=DEBUG uv run uvicorn src.ai_agent.main:app --reload --port 8080

# Run tests
uv run pytest tests/ -v

# Run tests with coverage
uv run pytest tests/ -v --cov=src/ai_agent --cov-report=xml
```

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

### Configure FastAPI Application Factory

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Shows how to use the `create_app` factory function to configure a FastAPI application. This includes setting up CORS middleware, exception handlers, and various routers for different API functionalities.

```python
from ai_agent.main import create_app, app

# Using the pre-configured app instance
# uvicorn src.ai_agent.main:app --reload

# Or create a new app instance
application = create_app()

# The app includes:
# - CORS middleware configured for all origins
# - Problem Details exception handlers
# - Health router at /api/v1/health
# - Metadata router at /api/v1/metadata/generate
# - Streaming router at /api/v1/metadata/generate-stream
# - Scalar docs at /scalar
# - Root endpoint at /

```

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

### POST /api/v1/metadata/generate

Source: https://github.com/aytymchuk/dilcore-ai-core/blob/main/README.md

Generates a structured JSON template from a given prompt using AI. This is a synchronous operation.

```APIDOC
## POST /api/v1/metadata/generate

### Description
Generate a template from a prompt (synchronous).

### Method
POST

### Endpoint
/api/v1/metadata/generate

### Parameters
#### Request Body
- **prompt** (string) - Required - The natural language prompt to generate the template from.

### Request Example
```json
{
  "prompt": "Create a contact form"
}
```

### Response
#### Success Response (200)
- **template** (object) - The generated JSON template.
- **explanation** (string) - An explanation of the generated template.

#### Response Example
```json
{
  "template": {
    "fields": [
      {
        "name": "name",
        "type": "text",
        "required": true
      },
      {
        "name": "email",
        "type": "email",
        "required": true
      },
      {
        "name": "message",
        "type": "textarea",
        "required": false
      }
    ]
  },
  "explanation": "This template represents a basic contact form with fields for name, email, and an optional message."
}
```
```

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

### Create and Serialize Template Response Model (Python)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Demonstrates creating a structured template response using Pydantic models, including nested models for sections and fields. It shows how to programmatically build a template object and then serialize it into a JSON string.

```python
from ai_agent.schemas.response import (
    TemplateResponse,
    TemplateSection,
    TemplateField,
    TemplateMetadata,
    TemplateStatus,
)

template = TemplateResponse(
    template_id="order-form-001",
    template_name="Order Form",
    description="E-commerce order submission template",
    status=TemplateStatus.ACTIVE,
    metadata=TemplateMetadata(
        version="2.0.0",
        author="Development Team",
        tags=["ecommerce", "orders", "forms"]
    ),
    sections=[
        TemplateSection(
            section_id="customer-info",
            title="Customer Information",
            description="Buyer details",
            fields=[
                TemplateField(
                    name="customer_email",
                    type="string",
                    required=True,
                    description="Customer email for order confirmation"
                ),
                TemplateField(
                    name="shipping_address",
                    type="object",
                    required=True,
                    description="Delivery address details"
                ),
                TemplateField(
                    name="quantity",
                    type="number",
                    required=True,
                    description="Number of items",
                    default_value=1
                )
            ]
        )
    ]
)

json_output = template.model_dump_json(indent=2)
```

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

### POST /api/v1/metadata/generate

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Generates metadata synchronously based on the provided input.

```APIDOC
## POST /api/v1/metadata/generate

### Description
This endpoint generates metadata synchronously. It takes a prompt and returns the generated metadata in a JSON response.

### Method
POST

### Endpoint
`/api/v1/metadata/generate`

### Parameters
#### Request Body
- **prompt** (string) - Required - The input prompt for metadata generation.

### Request Example
```json
{
  "prompt": "Create a form for user registration with fields for name, email, and password."
}
```

### Response
#### Success Response (200 OK)
- **generated_metadata** (object) - The generated metadata.

#### Response Example
```json
{
  "generated_metadata": {
    "form_schema": {
      "type": "object",
      "properties": {
        "name": {"type": "string"},
        "email": {"type": "string", "format": "email"},
        "password": {"type": "string", "format": "password"}
      },
      "required": ["name", "email", "password"]
    }
  }
}
```
```

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

### FastAPI Application Factory

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Information on creating and configuring the FastAPI application using the `create_app` factory function.

```APIDOC
## FastAPI Application Configuration

### Description
The `create_app()` function provides a factory pattern for setting up the FastAPI application, including essential middleware and routing.

### Key Configurations
- **CORS Middleware**: Configured to allow requests from all origins.
- **Exception Handlers**: Implements RFC 7807 Problem Details for standardized error responses.
- **Routers**: Includes endpoints for:
    - Health checks at `/api/v1/health`
    - Metadata generation (synchronous) at `/api/v1/metadata/generate`
    - Metadata generation (streaming) at `/api/v1/metadata/generate-stream`
- **API Documentation**: Scalar documentation is available at `/scalar`.
- **Root Endpoint**: A basic endpoint at `/`.

### Usage
```python
from ai_agent.main import create_app

application = create_app()
```

### Running the Application
Use `uvicorn` to run the application:
```bash
# Start development server
uvicorn src.ai_agent.main:app --reload

# Start on a custom port with debug logging
LOG_LEVEL=DEBUG uv run uvicorn src.ai_agent.main:app --reload --port 8080
```
```

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

### Handle AI Agent Exceptions in Python

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Demonstrates how to catch and handle custom AI agent exceptions in Python. It shows how to access error details like status code, problem type, title, and message from caught exceptions.

```python
from ai_agent.exceptions import AIAgentException, ValidationError, LLMProviderError, TemplateGenerationError, TemplateParsingError

# All exceptions include problem_type, title, status_code, and message
try:
    # Validation error (400)
    raise ValidationError("Prompt must be between 1 and 4000 characters")
except AIAgentException as e:
    print(f"Status: {e.status_code}")  # 400
    print(f"Type: {e.problem_type}")    # validation-error
    print(f"Title: {e.title}")          # Validation Error
    print(f"Detail: {e.message}")       # Prompt must be between 1 and 4000 characters

# LLM provider error (502)
raise LLMProviderError("Unable to communicate with AI provider")

# Template generation error (500)
raise TemplateGenerationError("An unexpected error occurred during template generation")

# Template parsing error (500)
raise TemplateParsingError("Unable to parse the generated template response")
```

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

### Stream Template Generation with Thinking Mode (Python)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Illustrates how to use the StreamingTemplateAgent for real-time template generation. It shows how to iterate over events, handling different event types like THINKING, CONTENT, TEMPLATE, ERROR, and DONE, to process streaming data.

```python
from ai_agent.agent import StreamingTemplateAgent, create_streaming_template_agent
from ai_agent.config import get_settings
from ai_agent.schemas.streaming import StreamEventType

settings = get_settings()
streaming_agent = create_streaming_template_agent(settings)

async for event in streaming_agent.generate_template_stream("Create a survey template"):
    if event.event_type == StreamEventType.THINKING:
        print(f"[Thinking] {event.data}")
    elif event.event_type == StreamEventType.CONTENT:
        print(f"[Content] {event.data}")
    elif event.event_type == StreamEventType.TEMPLATE:
        template_data = event.data
        print(f"[Template] {template_data['template']['template_name']}")
        print(f"[Explanation] {template_data['explanation']}")
    elif event.event_type == StreamEventType.ERROR:
        print(f"[Error] {event.data}")
    elif event.event_type == StreamEventType.DONE:
        print("Generation complete")
```

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

### Generate Template via SSE Stream (Bash)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

This snippet illustrates how to generate a template using Server-Sent Events (SSE) for real-time feedback. It sends a POST request to the streaming endpoint and receives events like 'thinking', 'content', 'template', 'error', and 'done' as the AI processes the request. The --no-buffer flag is used for immediate output.

```bash
# Stream template generation with real-time events
curl -X POST http://localhost:8000/api/v1/metadata/generate-stream \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Create a contact form template"}' \
  --no-buffer

# SSE event stream output
data: {"event_type":"content","data":"Analyzing your request...","timestamp":"2024-01-15T10:30:00Z"}

data: {"event_type":"content","data":"Creating template structure...","timestamp":"2024-01-15T10:30:01Z"}

data: {"event_type":"template","data":{"template":{"template_id":"contact-form","template_name":"Contact Form","description":"Template for contact submissions","status":"draft","metadata":{"version":"1.0.0","author":"AI Agent","tags":["contact","form"]},"sections":[{"section_id":"contact-info","title":"Contact Information","fields":[{"name":"name","type":"string","required":true},{"name":"email","type":"string","required":true},{"name":"message","type":"string","required":true}]}]},"explanation":"This template includes essential contact form fields organized in a single section."},"timestamp":"2024-01-15T10:30:02Z"}

data: {"event_type":"done","data":"Stream completed","timestamp":"2024-01-15T10:30:02Z"}
```

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

### Generate Template Synchronously (Bash)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

This snippet shows how to generate a structured JSON template synchronously using the AI Template Agent's API. It sends a POST request with a natural language prompt to the generation endpoint and receives a complete template in the response body. The prompt must be between 1 and 4000 characters.

```bash
# Generate a user registration template
curl -X POST http://localhost:8000/api/v1/metadata/generate \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Create a user registration form with email, password, and profile fields"}'

# Expected response
{
  "template_id": "user-registration-form",
  "template_name": "User Registration Form",
  "description": "Template for collecting user registration data including authentication and profile information",
  "status": "draft",
  "metadata": {
    "version": "1.0.0",
    "created_at": "2024-01-15T10:30:00Z",
    "author": "AI Agent",
    "tags": ["user", "registration", "form", "authentication"]
  },
  "sections": [
    {
      "section_id": "authentication",
      "title": "Authentication",
      "description": "User credentials for account access",
      "fields": [
        {
          "name": "email",
          "type": "string",
          "required": true,
          "description": "User email address for login",
          "default_value": null
        },
        {
          "name": "password",
          "type": "string",
          "required": true,
          "description": "Secure password for account",
          "default_value": null
        }
      ]
    },
    {
      "section_id": "profile",
      "title": "Profile Information",
      "description": "Personal details for user profile",
      "fields": [
        {
          "name": "full_name",
          "type": "string",
          "required": true,
          "description": "User's full name",
          "default_value": null
        },
        {
          "name": "bio",
          "type": "string",
          "required": false,
          "description": "Short user biography",
          "default_value": ""
        }
      ]
    }
  ]
}
```

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

### Generate Template (Synchronous)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Accepts a natural language prompt and returns a complete structured template with sections, fields, and metadata in a synchronous manner.

```APIDOC
## POST /api/v1/metadata/generate

### Description
Generates a structured JSON template based on a provided natural language prompt. The prompt must be between 1 and 4000 characters.

### Method
POST

### Endpoint
/api/v1/metadata/generate

### Parameters
#### Query Parameters
None

#### Request Body
- **prompt** (string) - Required - The natural language prompt describing the desired template.

### Request Example
```json
{
  "prompt": "Create a user registration form with email, password, and profile fields"
}
```

### Response
#### Success Response (200)
- **template_id** (string) - Unique identifier for the generated template.
- **template_name** (string) - The human-readable name of the template.
- **description** (string) - A brief description of the template's purpose.
- **status** (string) - The current status of the template (e.g., "draft").
- **metadata** (object) - Contains metadata about the template, such as version, creation timestamp, author, and tags.
  - **version** (string) - The version of the template.
  - **created_at** (string) - The timestamp when the template was created.
  - **author** (string) - The creator of the template.
  - **tags** (array of strings) - Keywords associated with the template.
- **sections** (array of objects) - An array of logical sections within the template.
  - **section_id** (string) - Unique identifier for the section.
  - **title** (string) - The title of the section.
  - **description** (string) - A description of the section's content.
  - **fields** (array of objects) - An array of fields within the section.
    - **name** (string) - The name of the field.
    - **type** (string) - The data type of the field (e.g., "string", "integer").
    - **required** (boolean) - Indicates if the field is mandatory.
    - **description** (string) - A description of the field.
    - **default_value** (any) - The default value for the field, if any.

#### Response Example
```json
{
  "template_id": "user-registration-form",
  "template_name": "User Registration Form",
  "description": "Template for collecting user registration data including authentication and profile information",
  "status": "draft",
  "metadata": {
    "version": "1.0.0",
    "created_at": "2024-01-15T10:30:00Z",
    "author": "AI Agent",
    "tags": ["user", "registration", "form", "authentication"]
  },
  "sections": [
    {
      "section_id": "authentication",
      "title": "Authentication",
      "description": "User credentials for account access",
      "fields": [
        {
          "name": "email",
          "type": "string",
          "required": true,
          "description": "User email address for login",
          "default_value": null
        },
        {
          "name": "password",
          "type": "string",
          "required": true,
          "description": "Secure password for account",
          "default_value": null
        }
      ]
    },
    {
      "section_id": "profile",
      "title": "Profile Information",
      "description": "Personal details for user profile",
      "fields": [
        {
          "name": "full_name",
          "type": "string",
          "required": true,
          "description": "User's full name",
          "default_value": null
        },
        {
          "name": "bio",
          "type": "string",
          "required": false,
          "description": "Short user biography",
          "default_value": ""
        }
      ]
    }
  ]
}
```
```

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

### POST /api/v1/metadata/generate-stream

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Generates metadata using streaming (Server-Sent Events) for real-time feedback.

```APIDOC
## POST /api/v1/metadata/generate-stream

### Description
This endpoint generates metadata using a streaming approach (Server-Sent Events). It's suitable for providing real-time updates during the generation process, useful for UI feedback.

### Method
POST

### Endpoint
`/api/v1/metadata/generate-stream`

### Parameters
#### Request Body
- **prompt** (string) - Required - The input prompt for metadata generation.

### Request Example
```json
{
  "prompt": "Generate a user profile template with fields for username, bio, and avatar URL."
}
```

### Response
#### Success Response (200 OK)
- **Server-Sent Events (SSE)** - The generated metadata is streamed in chunks.

#### Response Example (SSE format)
```
data: {"chunk": "Generating username field..."}

data: {"chunk": "Adding bio field..."}

data: {"chunk": "Finalizing avatar URL field."}

data: {"metadata": {"template": {"fields": [{"name": "username"}, {"name": "bio"}, {"name": "avatar_url"}]}}}
```
```

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

### Import Custom AI Agent Exceptions (Python)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Lists the custom exception classes provided by the AI Agent library. These exceptions are designed to map to Problem Details (RFC 7807) for consistent API error reporting.

```python
from ai_agent.exceptions import (
    AIAgentException,
    ValidationError,
    TemplateGenerationError,
    LLMProviderError,
    TemplateParsingError,
    ConfigurationError,
)
```

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

### Generate Template Stream (SSE)

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Provides real-time generation feedback using Server-Sent Events (SSE) as the AI processes the request.

```APIDOC
## POST /api/v1/metadata/generate-stream

### Description
Initiates a template generation process that streams results in real-time using Server-Sent Events (SSE). This endpoint emits various event types such as `thinking`, `content`, `template`, `error`, and `done`.

### Method
POST

### Endpoint
/api/v1/metadata/generate-stream

### Parameters
#### Query Parameters
None

#### Request Body
- **prompt** (string) - Required - The natural language prompt describing the desired template.

### Request Example
```bash
curl -X POST http://localhost:8000/api/v1/metadata/generate-stream \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Create a contact form template"}' \
  --no-buffer
```

### Response
#### Success Response (200) - SSE Stream
The response is a stream of Server-Sent Events. Each event has an `event_type` and `data` field.

- **event_type** (string) - The type of event being sent (e.g., `thinking`, `content`, `template`, `error`, `done`).
- **data** (any) - The payload of the event. The structure of `data` depends on the `event_type`.
  - For `content` events, `data` is a string message.
  - For `template` events, `data` is a JSON object representing the generated template (similar to the synchronous response).
  - For `error` events, `data` is an error object.
  - For `done` events, `data` is a completion message.

#### Response Example (SSE Stream Output)
```
data: {"event_type":"content","data":"Analyzing your request...","timestamp":"2024-01-15T10:30:00Z"}

data: {"event_type":"content","data":"Creating template structure...","timestamp":"2024-01-15T10:30:01Z"}

data: {"event_type":"template","data":{"template":{"template_id":"contact-form","template_name":"Contact Form","description":"Template for contact submissions","status":"draft","metadata":{"version":"1.0.0","author":"AI Agent","tags":["contact","form"]},"sections":[{"section_id":"contact-info","title":"Contact Information","fields":[{"name":"name","type":"string","required":true},{"name":"email","type":"string","required":true},{"name":"message","type":"string","required":true}]}]},"explanation":"This template includes essential contact form fields organized in a single section."},"timestamp":"2024-01-15T10:30:02Z"}

data: {"event_type":"done","data":"Stream completed","timestamp":"2024-01-15T10:30:02Z"}
```
```

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

### AI Agent Exceptions

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Details on the exceptions raised by the AI Agent, including their structure and common error types.

```APIDOC
## AI Agent Exception Handling

### Description
All exceptions raised by the AI Agent include a standard set of fields for consistent error reporting.

### Exception Fields
- **problem_type** (string) - A machine-readable identifier for the type of error.
- **title** (string) - A human-readable title for the error.
- **status_code** (integer) - The HTTP status code associated with the error.
- **message** (string) - A detailed message explaining the error.

### Common Error Types
- **ValidationError** (400 Bad Request): Occurs when input data fails validation, e.g., prompt length constraints.
- **LLMProviderError** (502 Bad Gateway): Indicates an issue communicating with the underlying LLM provider.
- **TemplateGenerationError** (500 Internal Server Error): An unexpected error during the template generation process.
- **TemplateParsingError** (500 Internal Server Error): Occurs when the generated template response cannot be parsed.
```

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

### Health Check Endpoint

Source: https://context7.com/aytymchuk/dilcore-ai-core/llms.txt

Verifies the service is running and returns configuration information, including the active AI model.

```APIDOC
## GET /api/v1/health

### Description
Checks the health status of the AI Template Agent service and retrieves its current configuration.

### Method
GET

### Endpoint
/api/v1/health

### Parameters
None

### Request Example
```bash
curl http://localhost:8000/api/v1/health
```

### Response
#### Success Response (200)
- **status** (string) - Indicates the health status of the service (e.g., "healthy").
- **app_name** (string) - The name of the application.
- **model** (string) - The currently active AI model being used.

#### Response Example
```json
{
  "status": "healthy",
  "app_name": "AI Template Agent",
  "model": "openai/gpt-oss-20b:free"
}
```
```

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

### POST /api/v1/metadata/generate-stream

Source: https://github.com/aytymchuk/dilcore-ai-core/blob/main/README.md

Streams the generation of a structured JSON template using Server-Sent Events (SSE). This allows for real-time updates as the template is being generated.

```APIDOC
## POST /api/v1/metadata/generate-stream

### Description
Stream template generation with SSE (real-time).

### Method
POST

### Endpoint
/api/v1/metadata/generate-stream

### Parameters
#### Request Body
- **prompt** (string) - Required - The natural language prompt to generate the template from.

### Request Example
```json
{
  "prompt": "Create a user registration form"
}
```

### Response
#### Success Response (200)
This endpoint returns a stream of Server-Sent Events (SSE).

**Event Types:**
- `thinking` - Reasoning content (if model supports thinking mode)
- `content` - Generation content chunks
- `template` - Final structured template with explanation
- `error` - Error event
- `done` - Stream completed

#### Response Example (SSE Stream)
```
event: thinking
data: {"content": "Model is thinking about the user registration form structure..."}

event: content
data: {"content": "Adding fields for username, email, and password..."}

event: template
data: {"template": {"fields": [{"name": "username", "type": "text", "required": true}, {"name": "email", "type": "email", "required": true}, {"name": "password", "type": "password", "required": true}]}, "explanation": "This template includes fields for username, email, and password for user registration."}

event: done
data: {"message": "Template generation complete."}
```
```

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