ROAPI

ROAPI automatically creates read-only APIs for static datasets without requiring any code. It builds on Apache Arrow and DataFusion: query frontends translate SQL, FlightSQL, GraphQL and REST requests into DataFusion plans, DataFusion executes them, a data layer loads datasets from many sources and formats with schema inference, and a response encoding layer serializes results into the format a client asks for.

You can point it at local files such as CSV or JSON, run it as a pre-built binary or Docker image, and query the data through a built-in web UI at /ui or with curl using SQL, GraphQL or REST. It can show the inferred schema for all tables, support MySQL and SQLite sources with table arguments, register data dynamically through an API, and read multiple table sources defined in YAML or TOML config files, including Google Sheets. Topics mention Parquet, S3 and Delta Lake.

ROAPI is written in Rust and licensed under Apache-2.0. It is aimed at slowly moving datasets and analytics use cases, and you run the server yourself.