> ## Documentation Index
> Fetch the complete documentation index at: https://koreai.mintlify.app/llms.txt
> Use this file to discover all available pages before exploring further.
# App Simulation and Testing
Test and validate your Agentic app interactively before deployment. Submit queries, observe agent responses in real time, and trace execution flow to catch issues early.
***
## Overview
The Playground is an interactive chat environment where you simulate conversations with your Agentic app. It captures a detailed execution timeline for each query and stores conversation history for debugging and review.
## Starting a Session
A new session starts automatically each time you open the Playground. To start a fresh session manually:
1. Click **New Session** at the top of the Playground interface.
2. The chat window clears and a blank session begins.
Starting a new session does not delete previous conversations. Access them using the **History** option.
## Interface Components
| Component | Description |
| ------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Query Input** | Text field for entering queries. Supports typed input and voice input via browser-based speech recognition (ASR). |
| **Response Display** | Shows the agent's response to each submitted query. When the [Artifacts is enabled](/agent-platform/agents/agentic-apps/settings/app-configurations#playground-settings), the response panel includes an artifacts section after each message. Artifacts are displayed only when the artifacts array contains valid entries, empty arrays are ignored. Playground supports AI for Service SDK templates. Hence, the rendering of artifacts depends on its template type field. If the type matches a supported template, the platform renders it as a native UI component, such as buttons, cards, or carousels, inline in the conversation. If the type is absent or unrecognized, the artifact is displayed as a formatted JSON object. [Learn More](/agent-platform/howto/structured-response-rendering-artifacts). |
| **Timeline** (Right Panel) | Step-by-step execution trace from user input to final output. Click a step to inspect details. |
| **Flow Diagram** (Right Panel) | Visual representation of the execution sequence across all components. |
| **Attachments** | Upload files as contextual input using the attach icon in the query input box. |
### Voice Input
Click the microphone icon in the query input field to activate voice input. Spoken input is converted to text and populated in the field automatically.
* Voice **output** is not supported — responses are text only.
* Voice input uses browser-based ASR, supported by default in Chrome, Firefox, Safari, and Edge.
### Timeline Details
Each step in the Timeline represents a distinct agent or tool (for example, Supervisor, Insights\_Agent) and shows the time taken to complete its task. Clicking a step reveals:
| Field | Content |
| -------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Messages** | The question or command sent to or received from the component, along with the reasoning used to process it. |
| **Request** | The prompt sent to the component. |
| **Response** | The output received from the component. |
| **Debug Logs** | Detailed status messages across agents, tools, and supervisors. Includes Guardrails execution logs, auto-expanded traces, and session and trace IDs. |
### Flow Diagram
The Flow Diagram provides a visual layout of the overall execution flow, making it easier to understand the sequence of operations and interactions between components.
### Attachments
Upload files to provide contextual information to the agent during a conversation. Platform uses attachments in two ways:
* Content Extraction - For specific file types, extracts and uses content as context.
* Metadata Storage - For all supported file types, stores file information for reference.
Learn More.
* Attachments require **Document Upload** to be enabled under Playground Settings on the Configurations page.
* File count and size limits are set under **Attachment Configurations** in General Settings. [Learn more](/agent-platform/agents/agentic-apps/settings/app-configurations#attachment-configuration).
* Uploaded files display a preview icon next to the file name for quick identification.
### Document Context Management
Uploaded documents are organized into two sections:
| Section | Description |
| -------------- | ------------------------------------------------------------------ |
| **In Context** | Documents actively used as context for the current conversation. |
| **Removed** | Documents that are uploaded but excluded from the current context. |
* **Remove from context**: Delete the document from **In Context** — it moves to **Removed** automatically.
* **Restore a document**: Click the **+** icon in the **Removed** section.
* **Delete permanently**: Click the delete icon on the file in the **Removed** section.
## Conversation History
The Playground automatically saves all conversation history. Use the **History** option to revisit, debug, or continue past sessions.
### Resuming a Session
1. Click **History** at the top of the Playground.
2. Browse past conversations — each entry shows the first user message and response. Use time filters to narrow the list.
3. Select a conversation to open it in the Playground.
4. Review or continue the conversation from where it left off.
### History Constraints
| Constraint | Detail |
| ------------------ | ---------------------------------------------------------------------------------------------------------------- |
| **Visibility** | Each user sees only their own conversation history. |
| **Message limit** | Only the most recent 50 messages (user inputs and agent responses) are stored and restored per session. |
| **Configuration** | Resumed sessions use the **current** configuration settings, not those from the original session. |
| **Attachments** | Attachments from the previous session remain accessible but must be manually re-selected to use them as context. |
| **Thought traces** | Available only for messages in the current active session — not for messages from resumed sessions. |
## Agent Tool Authorization
If an agent's tools use Auth Profiles that require user authorization, you are prompted at the start of the session.
| State | Behavior |
| ------------------ | --------------------------------------------------------------------------------------- |
| **Authorized** | Status shows "Authorized" and the agent proceeds with the query. |
| **Not Authorized** | A secure redirect URL is provided. Complete authorization before the agent can respond. |
The agent cannot execute actions until authorization is complete. If multiple authorizations are required, all are listed in a table at the start of the session. Unauthorized tools fail and return errors.
**Example**: If the agent requires access to Google services, the Playground displays the authorization status at the beginning of the session.
***