Skip to content
Nappai Documentation

BasicLangchainV1.0 Agent

The BasicLangchainV1.0 Agent is a conversational AI component designed to help you automate tasks and answer questions within the Nappai automation system. Think of it as a smart assistant that can not only chat with you but also perform specific actions or fetch external information when needed.

This component is built on the Langchain framework, which allows it to intelligently decide whether to answer a question directly using its training or to use external “tools” (like searching the web or accessing a database) to provide a more accurate and up-to-date answer.

How it Works

When you send a message or ask a question to this component, it follows a simple process:

  1. Understanding: The agent analyzes your input to understand what you are asking.
  2. Decision: It decides if it can answer the question using its internal knowledge or if it needs help from a specific tool (for example, checking a live website or a specific software application).
  3. Execution:
    • If it can answer directly, it generates a text response.
    • If it needs a tool, it uses that tool to gather information or perform an action, then combines that result with its knowledge to form a final answer.
  4. Response: Finally, it outputs the final text response back to the user or the next step in your automation workflow.

Essentially, it acts as a bridge between you and the various data sources or applications connected to your Nappai dashboard, making complex tasks feel like a simple conversation.

Connection & Credentials

This component does not require specific API keys or credential configurations in the Nappai panel before you can use its basic conversational features. However, if you plan to connect it to specific external tools (like a CRM or a database), you will need to ensure those tools are properly configured in your Nappai environment.

Inputs

The component relies on standard inputs typical for chat-based agents. While specific configuration fields are managed internally, you will interact with it primarily through:

  • User Message: The text or question you want the agent to answer.
  • Chat History: (Implicit) The agent remembers previous messages in the conversation to provide coherent answers, so context is automatically maintained during a session.

Outputs

The primary output of this component is a text response.

  • Final Answer: A string containing the agent’s reply to your query. This can be displayed to the user or passed to other components in your workflow for further processing.
  • Execution Logs: (Internal) The component may generate logs indicating whether it used a tool, if an error occurred, or if the process was successful. These logs are useful for troubleshooting but are typically not sent to the next step in the workflow.

Output Data Example (JSON)

Below is a simplified example of the structure of the text output you might receive. Note that the actual output is a text string, not always a complex JSON object, but it may contain information in this format if structured data is returned. json { “message”: “I have checked the latest inventory records for you. The item ‘Widget A’ is currently in stock.”, “used_tool”: false, “status”: “success” }

Connectivity

In a typical Nappai dashboard workflow, this component is usually connected as follows:

  • From: It often receives input from a Chat Interface or a Trigger component that captures user questions.
  • To: Its output is typically connected to a Response Sender (to show the answer to the user), a Text Processor (to refine the answer), or a Decision Node (to decide the next step in the automation based on the answer).

This node is versatile and can be placed anywhere in a workflow where an intelligent, conversational response is needed.

Usage Example

Scenario: You want to create a bot that answers customer questions about product availability.

  1. Add the BasicLangchainV1.0 Agent to your dashboard.
  2. Connect a “Get Customer Question” component to the input of the Agent.
  3. Configure any available “tools” (if supported by your Nappai setup) to point to your product database.
  4. When a customer asks, “Do we have the Blue Widget in stock?”, the Agent will:
    • Receive the question.
    • Decide to check the database (tool).
    • Receive the stock status from the database.
    • Generate a response: “Yes, we currently have 15 units of the Blue Widget in stock.”
  5. Connect the output of the Agent to a “Show Answer” component to display the result to the customer.

Important Notes

🟡 Development Status This component is currently marked as being in Development Mode. This means:

  • The functionality is fully usable for testing and building workflows.
  • However, the behavior or settings might change in future updates as the component is refined.
  • It is recommended to test thoroughly in a non-critical environment before using it for important, high-volume automated processes.

Tips and Best Practices

  • Keep Questions Clear: Since this is an AI agent, providing clear and specific questions will yield better results. Avoid vague queries.
  • Use Tools: If your automation requires up-to-date data (like live prices or stock levels), ensure that the relevant tools are connected and enabled for this agent to access them.
  • Monitor Conversations: If the agent gives unexpected answers, check the chat history input to ensure previous context is not confusing the AI.

Security Considerations

  • Data Privacy: Be mindful of what information you share with the agent. Since it may use external tools, ensure that any data it accesses complies with your organization’s security policies.
  • Development Mode: As this component is in development, avoid using it with sensitive production data until it has been fully validated and released as a stable version.

Templates

This section is omitted as this component is not currently part of any pre-defined templates.

This section is omitted as no specific related components are linked.