Skip to main content
This open-source script, available on GitHub at dust-tt/dust-labs/front automates the process of exporting your Front conversations and importing them into a Dust datasource. It preserves comprehensive conversation data and metadata, like so:
The script then formats and uploads this data to Dust, maintaining the conversation structure and metadata integrity.

​
How it Works

  1. Fetch Inboxes: Retrieves all available Front inboxes
  2. Process Conversations: For each inbox, fetches conversations (no time filtering applied)
  3. Fetch Messages: For each conversation, retrieves all messages
  4. Create Documents: Combines conversation metadata and messages into a single document
  5. Upload to Dust: Stores each conversation as a document with appropriate tags

​
Document Structure

Each conversation document contains:
  • Conversation Summary: Subject, status, assignee, timestamps, message count
  • Messages Section: All messages with sender, timestamp, direction, and content
  • Size Limits: Total document size limited to 2MB
  • Tags: Source, type, conversation ID, status, message count, timestamps, assignee, subject
This integration enables more effective use of Dust’s capabilities for creating AI agents that can analyze customer support responses, automatically draft new responses based on historical data, and summarize common support themes.