#2 · Primary category: AI Data Infrastructure & Storage

simdjson

aarch64 arm64 avx2 avx512 c-plus-plus clang clang-cl cpp11 gcc-compiler json json-parser json-pointer loongarch loongarch64 neon risc-v simd sse42 vs2019 x64

Parsing gigabytes of JSON per second : used by Facebook/Meta Velox, the Node.js runtime, ClickHouse, WatermelonDB, Apache Doris, Milvus, StarRocks

Project last updated:08/27/26

GitHub Stars

24.2K

Forks

1.3K

Contributors

195

License

Apache-2.0

Why we included this project

When a service has to chew through large JSON payloads, parsing often becomes the bottleneck, and simdjson exists to fix that. It uses SIMD instructions and microparallel algorithms to validate and parse JSON at gigabytes per second, several times faster than typical parsers, without skimping on full JSON and UTF-8 validation. That speed is why it's embedded in production systems like the Node.js runtime, ClickHouse, and Meta's Velox, and it fits naturally into AI data infrastructure where vector databases and analytics engines ingest huge JSON streams. The library drops into a C++ project with minimal fuss, picks a CPU-tailored parser at runtime with no configuration, and comes with solid documentation and bindings for other languages. For anyone building a database, a data pipeline, or any high-throughput service that parses JSON, this is a well-tested building block worth a look.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category