We are pleased to announce the public release of syNumpy, a standalone C++17 library from Symisc Systems for reading and writing NumPy .npy files.
syNumpy was built to make interoperability between C++ and Python/NumPy workflows simpler, especially in systems where numerical arrays, embeddings, feature vectors, and intermediate model outputs need to move cleanly between both environments.
The library is intentionally small and straightforward to integrate:
- single public header:
synumpy.hpp - single implementation file:
synumpy.cpp - modern C++17 implementation
- no external runtime dependency for the core library
- support for loading from disk or directly from memory buffers
- typed save helpers for common native C++ numeric types
The core parser entry point is syNumpy::loadNpyBuffer(), which makes the library useful in embedded, in-memory, or pipeline-driven workflows where the full file is already available as a byte buffer.
syNumpy is already used internally in production by FACEIO for facial feature extraction workflows, and by PixLab / Symisc Systems across internal computer vision, visual search, and document-processing systems.
Project links:
- Project page: https://pixlab.io/numpy-cpp-library
- Source code: https://github.com/symisc/sy-numpy-cpp
- FACEIO: https://faceio.net/
- PixLab: https://pixlab.io/
The library is released under the BSD 3-Clause License.
If your stack combines Python-based NumPy workflows with native C++ components, syNumpy may be a practical way to reduce friction between both sides.
Feedback, bug reports, and contributions are welcome.
