A high-performance C++ JSON parser that uses SIMD instructions to parse gigabytes of JSON per second.
simdjson is a high-performance C++ JSON parsing library that uses SIMD instructions and microparallel algorithms to parse JSON at gigabytes per second. It solves the performance bottleneck of JSON parsing in servers and data-intensive applications by offering full validation with unprecedented speed.
Developers and engineers working on high-performance servers, data processing pipelines, databases, and applications where JSON parsing is a critical performance factor, such as in Node.js, ClickHouse, and other data-intensive systems.
Developers choose simdjson for its unmatched parsing speed—4x faster than RapidJSON—while maintaining full JSON and UTF-8 validation, automatic CPU optimization, and easy integration with no configuration needed.
Parsing gigabytes of JSON per second : used by Facebook/Meta Velox, the Node.js runtime, ClickHouse, WatermelonDB, Apache Doris, Milvus, StarRocks
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Parses JSON at gigabytes per second—4x faster than RapidJSON and 25x faster than JSON for Modern C++—as validated by benchmarks on Intel Skylake processors.
Provides complete JSON and UTF-8 validation with lossless parsing, ensuring data integrity without compromising speed, as emphasized in its design philosophy.
Dynamically selects CPU-specific parsers at runtime with no configuration, leveraging SIMD instructions for optimal performance across different processors.
Includes record-breaking utilities like JSON minification at 6 GB/s, UTF-8 validation at 13 GB/s, and NDJSON processing at 3.5 GB/s, as listed in key features.
Requires modern C++ compilers (g++ 7+ or clang++ 6+) and a 64-bit system, limiting portability to older or non-standard environments, as noted in the quick start prerequisites.
Core functionality is C++-centric; while bindings exist for Python, Rust, and others, they add layers of complexity and may not offer full feature parity.
Optimized for large datasets, so the setup and SIMD-driven architecture might be excessive for applications parsing only tiny, occasional JSON files.