facebookincubator/velox

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

4201 stars · 1592 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Velox is a composable, extensible C++ execution engine library for data management systems, created by Meta. It provides reusable vectorized components—types, columnar vectors, expression evaluation, functions, operators, and I/O connectors—that developers use to build analytical query engines for batch, interactive, streaming, and AI/ML workloads.

Use cases

• build a custom analytical query engine in C++
• accelerate Apache Spark with native execution via Apache Gluten
• execute optimized query plans with vectorized operators
• read and process Parquet, ORC, and Nimble files from S3, HDFS, or GCS
• implement Presto or Spark compatible function semantics
• debug and replay query plan fragments in isolation
• build stream processing or AI/ML data pipelines on a shared execution engine

When to choose

• you are building or optimizing a compute engine and need high-performance vectorized execution primitives
• you want Arrow-compatible columnar processing with encodings like dictionary and RLE
• you need consistent Presto/Spark function semantics across multiple engines
• you want to offload Spark or Presto execution to a native C++ runtime

When to avoid

• you need an end-to-end SQL query engine with a parser and optimizer—Velox provides neither
• you are an end user looking for a ready-to-run database or analytics tool
• your project is not C++ or lacks integration effort budget for a low-level library

Facets

library · maturity active

database serialization data-science etl streaming databases big-data analytics cpp cross-platform query-execution columnar arrow-compatible vectorized-execution query-engine presto spark data-engineering linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="facebookincubator/velox")