Coding Prompt Cheat Sheet
For when you need to tell cursor or replit what a vCon is in a prompt...
Document for Coding Assistants: Structure of a vCon (Virtual Conversation)
Introduction
A vCon (Virtual Conversation) is a standardized JSON-based container designed to store and exchange data about real-time human conversations. It can encapsulate information from various communication modes, such as phone calls, video conferences, SMS, MMS, emails, and more. A vCon organizes this data into structured components for use in applications, data analysis, and regulatory compliance.
Key Components of a vCon
A vCon contains five main sections:
Metadata: Provides details about the conversation's context, including unique identifiers, timestamps, subject, and references to previous versions of the vCon.
Parties: Captures details about the participants in the conversation, including their roles, identifiers, and contact information.
Dialog: Stores the actual content of the conversation (e.g., text, audio, or video).
Analysis: Includes derived data such as transcripts, translations, sentiment analysis, or semantic tagging.
Attachments: Stores additional files related to the conversation, such as slides, images, or documents.
Structure of a vCon JSON Object
A vCon JSON object can be in one of three forms:
Unsigned: Initial or intermediate state during data collection.
Signed: Verified state with a digital signature for immutability.
Encrypted: Secure state to protect sensitive data.
Below is a breakdown of the JSON keys and their values:
General Structure
Metadata
vcon: Syntax version of the vCon.
uuid: A globally unique identifier for the vCon instance.
created_at: Timestamp for when the vCon was created.
updated_at: Timestamp for the last update to the vCon.
subject: Free-text field describing the topic of the conversation.
Parties
Each participant in the conversation is represented as an object:
Dialog
Each segment of the conversation is captured as a dialog object:
Analysis
Derived insights about the conversation are stored in analysis objects:
Attachments
Attachments provide supplemental data:
Security and Integrity
Signing: vCons can be signed using JWS (JSON Web Signature) to ensure their integrity and authenticity.
Encryption: Sensitive vCons can be encrypted using JWE (JSON Web Encryption) to protect their content.
Versioning: Redacted and appended vCons reference their original versions to maintain a history of changes.
Example Use Cases
Customer Support: Storing call recordings, transcripts, and attachments for quality assurance and analytics.
Legal Compliance: Maintaining immutable records of conversations with signatures for regulatory purposes.
Machine Learning: Using vCon data as input for training AI models while adhering to data privacy laws.
Conclusion
This document outlines the structure and components of a vCon to guide your coding team in implementation. Follow the described JSON schema to ensure compliance with the vCon standard for storing and exchanging conversation data.
Last updated
Was this helpful?