Skip to content
Nappai Documentation

Conversation Files

Conversation Files is a simple component that helps you bring conversation data into your Nappai workflows.
It lets you choose which conversation files you want to work with and passes that information to the rest of your automation.

How it Works

The component is built on top of Nappai’s ConversationMediaToolsBase.
When you drop it into a dashboard, you only need to provide three pieces of information:

  1. Embedding – a handle that tells the system where to find the conversation data.
  2. Model – the language model you want to use for any processing that follows.
  3. Tools – a list of tools you want to activate for this conversation (e.g., summarization, sentiment analysis, etc.).

Once the inputs are set, the component internally calls two helper methods:

  • fetch_conversation_files – pulls the requested conversation files from your data store.
  • build_conversation_media_tool – prepares the selected tools so they can be used later in the workflow.

No external APIs are called directly by this component; all work happens inside Nappai.

Inputs

  • Embedding: A handle that points to the conversation data you want to use.
  • Model: The language model that will process the conversation data.
  • Tools: A required list of tools you want to enable for this conversation. Select one or more from the dropdown.

Outputs

  • Data: The conversation files that were fetched. This output can be fed into other components that need the raw conversation content.
  • Tools: A set of tool objects built from the selected tools. These can be passed to downstream components that perform actions like summarizing or extracting insights.

Usage Example

  1. Drag the Conversation Files component onto your canvas.
  2. In the Embedding field, choose the conversation data source (e.g., “Customer Support Chat”).
  3. Select a language model in the Model field (e.g., “GPT‑4”).
  4. Pick the tools you need in the Tools field (e.g., “Summarizer”, “Sentiment Analyzer”).
  5. Connect the Data output to a component that displays or stores the conversation, and connect the Tools output to a component that runs the selected analyses.

The component will automatically fetch the conversation files and make the chosen tools available for the next steps in your workflow.

  • ConversationTool – Handles general conversation data operations.
  • FileTool – Manages file uploads, downloads, and storage.
  • MediaTool – Works with audio and video media within conversations.

Tips and Best Practices

  • Keep the Tools list focused; selecting too many tools can slow down the workflow.
  • Use a dedicated embedding for each conversation type (e.g., support chats vs. sales calls) to keep data organized.
  • Test the component with a small set of files first to confirm the correct data is being fetched.

Security Considerations

  • Ensure that the embedding you select only contains data you have permission to access.
  • The language model you choose may process sensitive information; use a model that complies with your organization’s privacy policies.
  • The component does not store data permanently; it only passes it to downstream components, so make sure those components handle data securely.