Cloud Parts

GizmoSQL

A database server that wraps DuckDB (or SQLite) and serves it to many clients over Arrow Flight SQL, with TLS and authentication.

What it is
Query engines
Type
Open source
Made by
GizmoData LLC
License
Apache 2.0 (Core); Enterprise features proprietary
Facts checked
6 October 2026
Website Source code

What it is

DuckDB runs inside one program, and only one process can write to a database file. GizmoSQL is a server process that owns that DuckDB instance and lets many clients share it over the network. Clients speak Arrow Flight SQL, a gRPC protocol in which results travel as Arrow record batches, so DuckDB’s columnar output reaches the client without being turned into rows. The server listens on port 31337 with TLS or mutual TLS, password login and server-signed JWT tokens. Connect with JDBC, ADBC (Python, Go, C/C++, R), ODBC or the bundled gizmosql_client shell; adapters cover dbt, SQLMesh, Ibis, SQLAlchemy, Superset, Metabase, Grafana and Power BI. Whatever DuckDB can do, the server can do: query Parquet on object storage, attach DuckLake catalogs, load extensions. SQLite is a second, switchable backend. It ships as Docker images, Homebrew and Windows installers, a Kubernetes operator and an iOS app meant for demos.

GizmoData LLC sells it as open core. GizmoSQL Core is Apache 2.0 and free. Enterprise adds session instrumentation and KILL SESSION, per-catalog permissions, SSO through OIDC, statement queuing and Prometheus metrics, unlocked by a signed licence key, at $300 per vCPU per year with a $1,500 yearly minimum per server. GizmoData Cloud, launched in September 2026, runs managed clusters on AWS, Google Cloud and Azure on prepaid plans from $29 a month. The current release, v1.41.0, bundles DuckDB 1.5.6, with 1.4.5 on a long-term support channel.

Why teams pick it

  • Turns a DuckDB file into a shared network database with TLS and logins, in one container
  • Columnar Arrow Flight SQL on the wire, with JDBC, ADBC and ODBC drivers so BI tools and notebooks connect directly
  • Full DuckDB SQL and extensions: Parquet and S3, DuckLake catalogs, spatial
  • Apache 2.0 core that runs on Linux, macOS, Windows and Kubernetes

Watch out for

  • Still one DuckDB on one machine: it scales up, not out
  • Operational controls such as SSO, per-catalog permissions, killing sessions, statement queuing and metrics are Enterprise only, billed per vCPU
  • The licence has moved before: it needed a licence key for production use until June 2025, and Enterprise code is proprietary
  • A small, founder-led vendor; Philip Moore wrote most of the commits
  • DuckDB now has its own server protocol, Quack; in the DuckDB team’s benchmark it moved 60 million rows in 4.9 s, against 17.4 s through GizmoSQL

Alternatives

What all query engines have in common →
Amazon AthenaAWSAWS’s serverless SQL engine for data in S3: no cluster to run, and on-demand queries cost $5 per TB scanned.Apache Arrow AceroExperimentalOpen sourceThe streaming execution engine inside Arrow’s C++ library: you hand it an exact plan of nodes (scan, filter, project, aggregate, join, sort) and it streams Arrow record batches through them, with no SQL parser or optimizer of its own.Apache CalciteOpen sourceA Java framework with everything a database needs to understand a query (SQL parser, validator, relational algebra and a rule-based, cost-based optimizer) but no storage of its own, which Hive, Flink, Druid, Dremio, Beam and many other systems embed as their SQL front end and planner.Apache DataFusionOpen sourceA query engine library in Rust, built on Arrow, that other projects embed and extend to build their own databases and data tools.+1 Apache DataFusion CometApache HadoopOpen sourceThe project that made big-data clusters normal: a distributed file system (HDFS), a cluster scheduler (YARN) and the MapReduce engine.Apache HiveOpen sourceThe first SQL engine for Hadoop, and the source of the Hive Metastore, the table catalog that Trino, Spark, Impala and Flink still read.Apache ImpalaOpen sourceAn MPP SQL engine in C++ for interactive queries over Hive tables on HDFS, Kudu, Iceberg and object storage; built and mainly sold by Cloudera.Apache PigDormantOpen sourceA step-by-step scripting language, Pig Latin, that compiles to MapReduce, Tez or Spark jobs on Hadoop. Still released, but the project calls itself dormant.Apache SparkOpen sourceThe general-purpose distributed engine for large batch jobs, streaming and machine learning, in Python, SQL, Scala, Java or R.Apache ZeppelinOpen sourceA self-hosted web notebook where each paragraph runs through an interpreter: Spark, Flink, JDBC, Python and more. Maintained, but behind Jupyter and Databricks in reach.chDBOpen sourceThe ClickHouse engine packaged as a library: run ClickHouse SQL inside a Python, Go, Rust or Node.js program, with no server.Cloudflare Basin SQLRenamedCommercialCloudflare’s serverless, distributed SQL engine for Apache Iceberg tables in Basin Catalog on R2, called R2 SQL until it became generally available on 1 October 2026, billed by the data it scans.DuckDBOpen sourceAn in-process analytical SQL database: a library with no server that queries Parquet, CSV, Iceberg and S3 directly.GlareDBDormantOpen sourceA small, embeddable SQL engine in Rust for querying Parquet and CSV files locally, over HTTP or in S3 and GCS, from a CLI, Python or the browser. It began as a DataFusion-based federated query service and was rewritten in 2025 on its own engine, and it has had no release since June 2025.HotdataCommercialA managed SQL engine for AI agents, built in Rust on Apache DataFusion and Arrow: each agent gets short-lived “instant databases” to load data into, then runs read-only SQL, vector, full-text and geospatial queries over an HTTP API or CLI.IbisOpen sourceA Python dataframe API that compiles to SQL, so the same code runs on DuckDB on a laptop and on BigQuery, Snowflake, PostgreSQL or more than 20 other engines.PolarsOpen sourceA DataFrame library with a Rust query engine underneath: you write Python (or Rust, SQL, Node.js, R) and it plans, optimises and runs the query on every core, streaming data bigger than memory.Presto (PrestoDB)Open sourceFacebook’s distributed SQL engine, now governed by the Linux Foundation, moving its workers from Java to C++ on Velox. Trino is its 2019 fork.RayforceOpen sourceA small embeddable engine in pure C that runs columnar analytics, graph traversals and recursive Datalog queries in one optimised plan.Spice.ai OSSOpen sourceA single-binary SQL engine in Rust, built on DataFusion and Arrow, that federates queries across databases, warehouses and lakes, keeps hot datasets in a local accelerated copy, and adds search and LLM APIs on top.TrinoOpen sourceA distributed SQL engine that queries data where it lives, from Iceberg tables on S3 to a live PostgreSQL database, in one query.VeloxOpen sourceMeta’s open-source C++ execution engine library: it takes an already-optimized query plan and runs it with vectorized operators, and it is the engine inside Presto’s C++ workers (Prestissimo) and Apache Gluten’s native Spark.

How it fits with other technologies

Serve a DuckDB file, then query it from another machine · bash

# Server: share ./data/sales.duckdb on port 31337 with TLS and a password
docker run --name gizmosql --detach --rm --tty --init \
  --publish 31337:31337 \
  --env TLS_ENABLED="1" \
  --env GIZMOSQL_PASSWORD="change-me" \
  --mount type=bind,source=$(pwd)/data,target=/opt/gizmosql/data \
  --env DATABASE_FILENAME="data/sales.duckdb" \
  gizmodata/gizmosql:latest

# Client: any Flight SQL driver works; this is the bundled shell (brew install gizmosql)
GIZMOSQL_PASSWORD="change-me" gizmosql_client \
  --host db.internal --username gizmosql_user --tls --tls-skip-verify \
  --command "SELECT region, sum(amount) AS revenue FROM orders GROUP BY region"

Editions and list prices#

List prices from GizmoData’s pages, checked on 6 October 2026, in US dollars.

GizmoSQL Core

free, Apache 2.0

The server, DuckDB and SQLite backends, Flight SQL, TLS and mTLS, passwords and JWT tokens, query timeouts. You run it.

GizmoSQL Enterprise

$300 per vCPU per year

Core plus SSO, per-catalog permissions, session instrumentation and KILL SESSION, statement queuing and Prometheus metrics. Minimum $1,500 a year per server. You run it, with a licence key.

GizmoData Cloud

from $29 a month

Managed GizmoSQL clusters on AWS, Google Cloud or Azure. The plan fee (Starter $29, Team $99, Business $499 a month) comes back as compute credit billed by the hour; managed storage is $0.05 per GB-month.

Three ways to share one DuckDB#

DuckDB itself runs inside one program. These are the main ways to let many clients use the same database.

GizmoSQLDuckDB QuackMotherDuck
What runsYour own server process holding DuckDBA DuckDB extension; any DuckDB can act as server or clientA managed service on AWS, one DuckDB instance per user
ProtocolArrow Flight SQL over gRPCQuack, over HTTPMotherDuck’s own protocol, plus a Postgres wire endpoint
ClientsJDBC, ADBC and ODBC drivers, BI toolsOther DuckDB instances, through ATTACH 'quack:host'DuckDB clients with ATTACH 'md:', Postgres drivers
CostCore free; Enterprise $300 per vCPU per yearFree, MITFree Lite plan; Business $250 a month per organisation plus usage

See it in a real stack

No architecture uses GizmoSQL yet. Drop it into the playground and wire it up yourself.Open playground →

Benchmarks

Benchmark · ClickHouse Inc.ClickBenchHow fast analytical databases run 43 SQL queries over one 100-million-row table of web analytics events, on the same cloud machines.Includes GizmoSQL · 7 caveats to read first
How to read a benchmark →
Where it came from2022–2026

Arrow Flight SQL gave databases a way to send query results as columnar Arrow data instead of row by row. In April 2022 Tom Drabas at Voltron Data started an example Flight SQL server with SQLite and DuckDB backends. Philip Moore, a long-time data engineer then at Voltron Data, did most of the later work, and the example grew into a project called SQLFlite.

DuckDB’s in-process design left no built-in way for many users to share one database over a network. In October 2024 Moore forked SQLFlite under his own company, GizmoData, as GizmoSQL. At first it was a licensed product that printed a warning when run without a key. In June 2025 he made it open source under Apache 2.0, and in January 2026 added a separate, proprietary Enterprise edition. GizmoData Cloud, a managed version, followed in September 2026. By then DuckDB had shipped its own client-server protocol, Quack, citing Flight SQL servers among the workarounds people had built.

Why it’s built this way

Arrow Flight SQL instead of the PostgreSQL wire protocol
DuckDB already produces Arrow data. Sending Arrow batches over gRPC keeps results columnar end to end, and the Arrow project supplies JDBC and ADBC drivers.
One server process that owns the database
Only one process can write to a DuckDB file. A server that holds the single instance lets many clients read and write through it.
Open core, with paid features in a separate directory
Core stays Apache 2.0, while enterprise controls live in src/enterprise/ under a commercial licence and unlock with a signed key.

Timeline

  1. 2022

    Voltron Data starts a Flight SQL server example on DuckDB and SQLite

  2. 2024

    Philip Moore forks it as GizmoSQL in October, as a licensed product

  3. 2025

    GizmoSQL becomes open source under Apache 2.0 in June

  4. 2026

    Proprietary Enterprise edition in January; DuckDB ships Quack in May

  5. 2026

    GizmoData Cloud launches in September

Sources20 · facts checked 6 October 2026

Improve this page on GitHub