> ## 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.

# Test (Talk to Bot)

<Badge icon="arrow-left" color="gray">[Back to NLP Topics](/ai-for-service/automation/natural-language/nlp-topics)</Badge>

Talk to Bot is a real-time chat interface for testing your app as an end user. Use it to validate recognition, NLP performance, and conversation flow.

<Note>
  The Test icon is not available until at least one task is created.
</Note>

***

## Start a Test Session

1. Click the **Test** icon on the top-right of the platform.
2. Select **Voice** or **Chat** (this article covers Chat).
3. Enter a message in the **Message** field to begin. Example: Book a flight.
4. The NLP interpreter processes the input, authenticates, prompts for required fields, and executes the task.

Try varied phrasings and synonyms to validate NLP coverage. If the app returns unexpected results, add or update synonyms and patterns. See [Natural Language Processing](/ai-for-service/automation/natural-language/training/optimizing-bots#optimizing-bots).

***

## Debug Window

Click the **Debug** icon on the top-right of the chat window to open the debug panel. It has three tabs:

### Debug Log

Shows the sequential execution of dialog task components (Script Node, Service Node, Webhook Node) with timestamps. Grouped by user inputs and app responses.

**Statuses:**

| Status                     | Description                                             |
| -------------------------- | ------------------------------------------------------- |
| `initiated`                | Node execution started (script, service, webhook).      |
| `execution`                | Node execution in progress.                             |
| `execution successful`     | Node completed successfully.                            |
| `process completed`        | Full execution process for the node is done.            |
| `parsing`                  | Platform is parsing the user prompt.                    |
| `parsing successful`       | Prompt parsed successfully.                             |
| `waitingForUserInput`      | User was prompted for input.                            |
| `pause`                    | Current dialog paused while another task runs.          |
| `resume`                   | Paused dialog continues after the other task completes. |
| `waitingForServerResponse` | Pending async server response.                          |
| `error`                    | Error occurred (loop limit, server/script failure).     |
| `end`                      | Dialog reached the end of the flow.                     |

Expand any node entry and click **JSON** to view detailed log data. You can copy or view the script in full-screen.

### NL Analysis

Shows intent detection and recognition scores for each utterance. Covers tone analysis, intent detection, and entity detection across ML, FM, and KG engines. See [Utterance Testing](/ai-for-service/automation/testing/testing-your-bot-with-nlp#utterance-testing) for score details.

### Session Context and Variables

Displays the `Context` object and session variables updated at each dialog component. See [Using Session and Context Variables](/ai-for-service/automation/agent-flows#session-management) and [Context Object](/ai-for-service/automation/agent-flows#context-object).

***

## System Commands

System Commands let you control the conversation during evaluation. They can also be injected via JavaScript. See [System Commands](/ai-for-service/automation/testing/talk-to-bot#system-commands).

***

## Record Session

Use the **Record** option to capture conversations for regression testing.

1. Click **Record** to start capturing the session. A notification appears at the top of the chat window.
2. Click **STOP** to end the recording.
3. After stopping, click **Create Test Case** to save the conversation as a conversation test case.
4. Optionally, download the recording as a JSON file.