ClickHouse vs DuckDB
Two analytics (OLAP) options for database. When each fits, what it costs, who moves from one to the other, and what makers who chose it say.
Ask your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗
Which fits you
Choose ClickHouse if- Fast aggregate queries over billions of rows of events, logs or metrics, self-hosted or on ClickHouse Cloud.
Use it whenDashboards or usage analytics are getting slow on your main Postgres.
Trade-offBuilt for bulk inserts and scans, not frequent row updates or transactions; runs next to your main database, not instead of it.
Choose DuckDB if- Analytical SQL inside your app or script over local or S3-hosted CSV and Parquet files, with no server to run.
Use it whenOne process does the analysis — a report job, a data notebook, analytics in the browser via WebAssembly.
Trade-offAn embedded library, so it doesn't serve many concurrent writers like a database server does.
At a glance
| ClickHouse | DuckDB | |
|---|---|---|
| Used by | 40 makers' products · 58 open-source projects | 14 makers' products · 44 open-source projects |
| Downloads | 3.1M/wk+41% vs npm | 1.2M/wk4.8× vs npm |
| Pricing | Free to self-host (Apache 2.0); ClickHouse Cloud is usage-based with compute and storage billed separately, starting with a 30-day trial and $300 in credits. · paid from $53/mo (Basic) | Free and open source (MIT). |
| Free tier | Yes | Yes |
| Open source | Yes · self-hostable | Yes |
| Incidents, 90 daysfrom its status page | 17 (2 major) | no public status feed |
What makers say
Makers on using it for database, from Product Hunt and Starter Story interviews, each linked to the source. Products with a page of their own and fuller notes first.
On ClickHouseWe built on ClickHouse to process large volumes of AI bot and server-log data in real time. Agent analytics requires speed and scale - and ClickHouse delivers both.
ClickHouse backs Basedash Warehouse, every team's centralized warehouse in Basedash. It's incredibly fast for complex queries and scales to any size we need.
Spans are append-only and every screen in the product is an aggregate over millions of them, which is a columnar problem, so that's where they live.
On DuckDBOur built-in SQL cells are powered by duckdb, which makes it possible to execute blazingly fast queries against Python dataframes as well as external databases like SQLite and Postgres.
We use DuckDB WebAssembly as the core of Universal SQL and it has made it possible to deliver incredibly fast, interactive apps. The DuckDB team and community are phenomenal.
DuckDB powers our analytics engine for SQLite, CSV, Excel files and more. We couldn't have built CamelAI without DuckDB. It's enabled us to execute SQL queries that used to take minutes in seconds.
Loved and watch-outs
Themes that recur in makers' words and Hacker News comments, each linked to what it summarises.
ClickHouseOn Product Hunt: 5.0★, 28 reviews · mentioned most: fast performance, real-time analyticsDuckDBOn Product Hunt: 5.0★, 16 reviews · mentioned most: fast performance, in-process SQL engine, WebAssembly supportWho uses each
Incerto
Prefactor
Siteline
Baselight
Databox
Evidence
Fabi.ai
marimo
datajoiWhat makers pair each with
With ClickHouseTransactional Email
ClickHouse + Resend:
BasedashBeam Analytics
FlexpriceGraphJSON
Laminar
Rybbit
Human Behavior
Sleek Analytics
Dodo Payments
Helicone
Blacksmith
Infisicaland 2 more
BasedashBeam AnalyticsFlexpriceGraphJSONLaminarRybbitHuman BehaviorSleek AnalyticsDodo PaymentsHeliconeBlacksmithInfisicaland 2 more
Render
Retool
Phare
Lago
SuperwallHosting
ClickHouse + Amazon Web Services:
DatazipFlexpriceLaminarPulse for Elasticsearch and OpenSearchHuman BehaviorDodo Payments
Langfuse
ArcadeBlacksmith
Fillout
Frill
GrowthBookand 7 more
DatazipFlexpriceLaminarPulse for Elasticsearch and OpenSearchHuman BehaviorDodo PaymentsLangfuseArcadeBlacksmithFilloutFrillGrowthBookand 7 moreFrontend Framework
Payments
ClickHouse + Stripe:
ASOBasedashBeam AnalyticsRybbitHuman BehaviorSleek AnalyticsHeliconeLangfuseArcadeFilloutFrillGrowthBookand 5 more
ASOBasedashBeam AnalyticsRybbitHuman BehaviorSleek AnalyticsHeliconeLangfuseArcadeFilloutFrillGrowthBookand 5 moreClickHouse + PayPal:
UptimeRobot
UptimeRobotWith DuckDBNeither quite right?
See the whole database decision →
SupabaseBackend platformPostgres plus auth, storage and realtime in one project.
ConvexBackend platformRealtime, collaborative apps written entirely in TypeScript.
FirebaseBackend platformMobile apps that need offline sync and client-side SDKs.
NeonManaged databaseServerless Postgres that scales to zero, with a branch per preview deploy.
PlanetScaleManaged databaseApps that expect heavy scale and want zero-downtime schema changes.
TursoManaged databaseSQLite at the edge, or one database per customer.
MongoDBManaged databaseDocument-shaped data and teams already fluent in Mongo.
PostgreSQLRun it yourselfThe engine behind most of the hosted options above, run on your own server or container for full control and a flat cost.