Rust-native, Spark-compatible data and AI platform that runs existing Spark code faster and cheaper with a built-in agent layer.
What it does
LakeSail is a managed data and AI platform built on a Rust engine that stays compatible with the Apache Spark Connect protocol. Existing PySpark, Spark SQL, Delta Lake and Iceberg code runs unchanged while gaining lower cost and higher speed, and it ships an MCP server and agent tooling from the start.
How to use: LakeSail can be used as a drop-in replacement for Spark SQL and the Spark DataFrame API. Installation is done via pip (pysail[spark]). A Sail server can be started and connected to from PySpark with minimal code changes. Kubernetes deployment is also supported.
Core features
Rust engine with full Spark API compatibility
Runs in your own AWS account (BYOC)
Iceberg and Delta Lake native support
Native Python and AI workloads without JVM overhead
MCP server and agent layer
Lakehouse branching and sub-second cold starts
Best for
→Migrating Spark workloads without rewrites
→Lowering data infrastructure costs
→Running AI agents against a data lakehouse
→Interactive notebook analytics
Toolspool rankingby monthly traffic