App¶

The App class is the top-level container for workflows in the Virtualitics AI Platform.

App Class¶

App ¶

App(name: str, description: str, image_path: Optional[str] = None, is_shareable: bool = False, sort_key: Optional[str] = None, agent: Optional[DispatcherAgentInterface] = None, iris_config: Optional[IrisConfig] = None)

An App represent a workflow in the Virtualitics AI Platform. An App can contain many :class:~virtualitics_sdk.flow.step.Step's that can get inputs, run computation, and show results.

Parameters:

name (str) –

The name of the App.
description (str) –

The App description
image_path (Optional[str], default: None ) –

The base64 encoding of an image or a public link to an image, defaults to None
is_shareable (bool, default: False ) –

Whether a given App is shareable, defaults to False
sort_key (Optional[str], default: None ) –

An optional value to sort the App by, this controls the sort order of Apps on the home page

agent (Optional[DispatcherAgentInterface], default: None ) –

EXAMPLE:

# Imports from virtualitics_sdk import App...
# Example usage tile_image_link = "https://example-image.jpeg" example_flow = App(name="Example App", description="Simple example for using App class", image_path=tile_image_link)

chain ¶

chain(ordered_steps: List[Step], lock: bool = True)

Add all steps to the App's workflow with one function. Locks Apps by default.

Parameters:

ordered_steps (List[Step]) –

Ordered list of Steps to be added.

Flow Metadata¶

FlowMetadata ¶

Usage Example¶

from virtualitics_sdk import App, Step, StepType, Page, Section

# Create an app
my_app = App(
    name="My Application",
    description="An example application",
    is_shareable=True
)

# Add steps using chain()
my_app.chain([
    step1,
    step2,
    step3
], lock=True)

Key Methods¶

chain()¶

The chain() method is the primary way to add steps to your app. It takes a list of steps in execution order and locks the app by default.

app.chain([step1, step2, step3])

SingletonApp¶

SingletonApp is a variant of App that ensures only one active instance exists at a time. Useful for monitoring dashboards and shared resources.

SingletonApp ¶

SingletonApp(name: str, description: str, image_path: Optional[str] = None, is_shareable: bool = False, sort_key: Optional[str] = None, agent: Optional[DispatcherAgentInterface] = None, trigger: Optional[Trigger] = None, iris_config: Optional[IrisConfig] = None)

from virtualitics_sdk import SingletonApp

app = SingletonApp(
    name="System Monitor",
    description="Live monitoring dashboard — single instance only"
)
app.chain([monitor_step])

When to Use SingletonApp¶

Live dashboards where multiple instances would be redundant
Shared resource monitors (database status, cluster health)
Admin tools that should only run once at a time

LLM Configuration¶

Apps can configure IRIS integration with hooks and default prompts:

async def on_llm_request(request_data, link, flow_metadata):
    """Add context before sending to LLM."""
    request_data["context"] = {"app_data": "..."}
    return request_data

app = App(
    name="AI Analysis",
    description="AI-powered analysis",
    on_llm_request=on_llm_request,
    default_prompts=[
        "What are the key trends?",
        "Summarize the data"
    ]
)

See the LLM Integration guide for details.

Best Practices¶

Give descriptive names: App names appear in the UI, make them clear
Set descriptions: Help users understand what your app does
Use shareable flag: Enable sharing if users should be able to share instances
Order matters: Steps execute in the order they're added via chain()
Lock your app: The lock=True parameter (default) prevents accidental modifications
Use SingletonApp for dashboards and monitors that shouldn't have multiple instances