Skip to content

Changelog

This document provides a chronological list of notable changes for each version of EncypherAI.

2.0.0 (04-13-2025)

Added

  • Ed25519 digital signatures for enhanced security
  • Key management utilities for generating and managing key pairs
  • Public key resolver pattern for verification
  • Improved API for metadata embedding and verification
  • Updated documentation with digital signature examples
  • C2PA-inspired manifest structure for enhanced content provenance
  • Interoperability module (encypher.interop.c2pa) for conversion between EncypherAI and C2PA-like structures
  • Comprehensive documentation on C2PA relationship and alignment

Changed

  • Replaced HMAC verification with Ed25519 digital signatures
  • Updated UnicodeMetadata class to be the primary interface
  • Deprecated MetadataEncoder and StreamingMetadataEncoder classes
  • Improved StreamingHandler to use digital signatures
  • Updated all examples and integration guides
  • Aligned manifest field names with C2PA terminology:
  • Renamed actions to assertions in ManifestPayload
  • Renamed action to label in ManifestAction
  • Renamed ai_info to ai_assertion in ManifestPayload
  • Updated documentation to reflect terminology changes
  • Enhanced docstrings with references to C2PA concepts

Security

  • Enhanced security with asymmetric cryptography
  • Separate private keys for signing and public keys for verification
  • Key ID system for managing multiple keys
  • Improved tamper detection capabilities

Documentation

  • Added new user guide: Relationship to C2PA Standards
  • Updated examples to use the new field names
  • Added code examples for the interoperability module

1.0.0 (03-22-2025)

Added

  • Initial stable release of EncypherAI
  • Core metadata encoding and decoding functionality
  • HMAC verification for tamper detection
  • Support for multiple embedding targets:
  • Whitespace (default)
  • Punctuation
  • First letter of words
  • Last letter of words
  • All characters
  • Streaming support for handling content from LLMs
  • Integration with popular LLM providers:
  • OpenAI
  • Anthropic
  • LiteLLM
  • Comprehensive documentation and examples
  • Interactive demos:
  • Jupyter Notebook demo
  • Streamlit web app
  • FastAPI example application
  • Python client library
  • JavaScript client library

Security

  • Secure HMAC verification using SHA-256
  • Secret key management for verification
  • Tamper detection capabilities