Skip to main content

Introduction

µQuery (micro query) is a lightweight HTTP API for querying data files using SQL, powered by DuckDB and built in Rust.

Send a SQL query as a plain HTTP request, get results back in the format you need — JSON, CSV, JSON Lines, or Arrow IPC. µQuery handles the rest.

Key features​

  • Format-agnostic: Query CSV, JSON, Parquet, Excel, and many more with plain SQL
  • Flexible output: Stream results as JSON, CSV, JSON Lines, or Apache Arrow IPC — negotiated via the Accept header
  • Cloud-native: Query files directly from S3, GCS, or HTTPS without downloading them first
  • Serverless-ready: Runs on AWS Lambda and Google Cloud Run with no infrastructure changes
  • Fast startup: Rust binary with pre-loaded DuckDB extensions — cold starts in milliseconds
  • Connection pooling: Concurrent queries handled via a configurable pool of DuckDB connections

How it works​

µQuery exposes a single POST / endpoint. The request body is a SQL query; the Accept header controls the response format:

curl -X POST http://localhost:8080 \
  -H "Content-Type: text/plain" \
  -H "Accept: application/vnd.apache.arrow.stream" \
  -d "SELECT * FROM 'https://example.com/data.parquet' LIMIT 100"

Results are streamed incrementally — the response starts as soon as the first batch is ready, so large queries don't require waiting for full materialization.