EQBook is built on exceptional open-source software. Full attributions and license texts are provided below.
Apple's array framework for machine learning on Apple Silicon. Powers all LLM inference in EQBook.
LLM generation utilities built on MLX. Used for text generation with Gemma 4 models.
Swift bindings for the MLX framework, enabling native Swift integration of on-device ML inference.
Transformer model implementations in Swift/MLX. Provides the Gemma architecture implementation used by EQBook.
High-performance C++ implementation of OpenAI's Whisper with CoreML acceleration. Powers on-device audio and video transcription in EQBook (Whisper Small model).
A vector search SQLite extension. Used as EQBook's default on-device vector store for semantic search and RAG retrieval.
Swift client for the Qdrant vector database. Used as an optional high-performance vector backend on macOS when Qdrant is running locally.
Sentence embedding model by sentence-transformers. Used to generate 384-dimensional vector embeddings for document chunks and queries in EQBook's RAG pipeline.
EQBook uses quantized Gemma 4 model weights (gemma-4-e2b-it-4bit, gemma-4-e4b-it-4bit, gemma-4-26b-a4b-it-4bit) from the mlx-community organisation on HuggingFace. These weights are governed by Google's Gemma Terms of Use, not the MIT License.
Key points of the Gemma Terms of Use:
The full Gemma Terms of Use are available at ai.google.dev/gemma/terms. EQBook downloads these weights on first launch; by proceeding with the download you agree to the Gemma Terms of Use.
Apple's Swift package for parsing, building, editing, and analysing Markdown documents. Used to render AI responses in EQBook's chat interface.
Encoder & decoder for XML using Swift's Codable protocols. Used to parse the XML payload inside .docx files (Office Open XML format) during source ingestion.
Effortless ZIP file handling in Swift. Used to unzip .docx and .epub containers before XML/HTML extraction during source ingestion.
Syntax highlighting for the web. Used to highlight code blocks in AI responses within EQBook's chat interface.
EQBook also uses the following Apple-provided system frameworks, which are subject to Apple's developer agreements and are not redistributed by EQBook:
All trademarks are the property of their respective owners. "Gemma" is a trademark of Google LLC. "Apple", "macOS", "Swift", "CoreML", and "Apple Silicon" are trademarks of Apple Inc. EQBook is an independent product and is not affiliated with, endorsed by, or sponsored by Apple Inc. or Google LLC.
If you believe a license attribution is missing or incorrect, please contact us at info@libardoramirez.com.