EncypherAI Documentation
Invisible Metadata for AI-Generated Text
Welcome to EncypherAI
EncypherAI is an open-source Python package that enables invisible metadata embedding in AI-generated text using zero-width characters.
This documentation provides comprehensive information about the EncypherAI package, including installation instructions, usage examples, API reference, and integration guides.
Key Features
| Feature | Description |
|---|---|
| 🔍 Invisible Embedding | Add metadata without changing visible content. |
| 🔐 Digital Signature Verification | Ensure data integrity and detect tampering with Ed25519 signatures. |
| 🏛️ C2PA-Inspired Manifests | Embed structured, verifiable manifests for text provenance. |
| 🌊 Streaming Support | Compatible with chunk-by-chunk streaming from LLMs. |
| 🧩 Modular Architecture | Clean separation of key management, payload handling, and signing. |
| 🔄 Extensible API | Easily integrate with any LLM provider. |
Demo Video
Watch our demo video to see EncypherAI in action
Getting Started
User Guide
- Basic Usage
- Metadata Encoding
- Extraction and Verification
- Tamper Detection
- Streaming Support
- C2PA Relationship
- Migration Guide
Examples
API Reference
UnicodeMetadata: Core class for embedding and verification.c2pa_interop: Utilities for C2PA compatibility.- Deprecated Classes: Info on legacy classes.
Contributing
We welcome contributions to EncypherAI! Check out our contribution guidelines to get started.
License
EncypherAI is provided under a dual licensing model designed to support both open-source community use via an AGPL-3.0 License and commercial enterprise adoption via commercial licenses.
GitHub Repository
Visit our GitHub repository to access the source code, report issues, or contribute to the project.