We’ve just released support to directly query the underlying ClickHouse database in Tinybird via the native ClickHouse HTTP interface. This makes it much easier to connect BI tools, SQL clients, and ORMs directly to your Tinybird workspaces using their built-in ClickHouse drivers.
This new interface will replace our now deprecated Postgres-based BI connector with a more effective, future-proof, and maintainable way of integrating Tinybird into your analytics stack.
Why we built this
ClickHouse is a production-tested OLAP database that has proven itself at massive scale, but it’s also relatively new. When we started Tinybird, BI and visualization tools didn’t yet offer first-party support simply because the ecosystem was still catching up. As ClickHouse has matured and grown in adoption, that has changed.
In the past, we bridged the ClickHouse-to-BI gap with a Postgres-based BI connector. It used a Postgres foreign data wrapper to connect and push queries down to ClickHouse under the hood, which had some benefits (every BI under the sun connects to Postgres) but also some inherent problems (it didn't always push down queries) which made it very inefficient in some cases.
With native ClickHouse support now widespread, it’s time to retire that connector and give our users a more direct and effective way to connect.
We also wanted to make sure we could expose access to the underlying ClickHouse database with the same guarantees of high availability, performance, load balancing, and security that Tinybird users get when querying via our API endpoints. And with this new interface, that's what we have done.
What this means for Tinybird users
If you’re using a BI tool or ORM with ClickHouse support, you can now connect directly to Tinybird without any extra plugins or workarounds. This makes it easier to do things like:
- Build observability dashboards in Grafana
- Explore data with Hex
- Connect your BI tool of choice like Superset or Metabase
- Use highly popular query editors like DBeaver
- Connect to Tinybird using ClickHouse's standard clients for Javascript, Python, Go or other languages.
What's new
The ClickHouse HTTP interface is built on top of Tinybird’s existing /v0/sql endpoint, extending it for compatibility with the particular requirements of most common BI tools and libraries. Here’s what you get in addition to being able to query:
System tables for introspection:
system.databases,system.tables,system.columns,system.partsand similar metadata tables that are native to ClickHouse are now available to query, both via the Tinybird UI and the new ClickHouse interface.Full Tinybird observability over queries made through the ClickHouse interface. They will show up in real-time in Service Data Sources (
tinybird.pipe_stats_rt,organization.pipe_stats_rt) as Query API requests. This gives you visibility into how your BI tools are querying your data, how much CPU these queries consume, duration, etc.Pipes as Views: Tinybird pipes are exposed in system.tables as
View(orMaterializedViewfor materializations). This makes pipes reusable in BI tools, just like tables.Shared Data Sources: In Tinybird you can share Data Sources between different workspaces. These will show up in
system.databases,system.tables, andsystem.columns, enabling consistent metadata visibility across workspaces.ClickHouse-compatible parameters: Parameters like
user,query, anddatabaseare supported, ensuring tools "just work" without awkward configuration.Authentication & headers: Use Tinybird auth tokens for authentication. Everything integrates cleanly with BI tool expectations.
tinybird-localsupport: The interface works with tinybird-local, too. Instead of subdomains (like in Cloud), the local setup uses a different port.Native integration with BI tools: As mentioned above, user can easily now connect to Grafana, Hex, Superset, Metabase, and more directly with existing ClickHouse connectors.
Getting started
To connect your BI tool, ORM, or language client to Tinybird via the ClickHouse native HTTP interface, simply prepend clickhouse. to the hostname for your workspace's Tinybird region. For example:
Protocol: HTTPS
Host: clickhouse.<REGION>.tinybird.co
Port: 443 (HTTPS)
SSL/TLS: Required (enabled)
Username: <WORKSPACE_NAME> # Optional, for identification purposes
Password: <TOKEN> # Your Tinybird auth token
Replace <REGION> with your workspace region (e.g., eu-west-1.aws), which you can get from Tinybird Cloud or via the Tinybird CLI. Check out the docs for the full list of supported regions, as well as all available connection parameters.
Example: Connect Grafana to Tinybird in 2 minutes
- In Grafana, add a new data source
- Select ClickHouse
- Set the Server URL to
https://clickhouse.<REGION>.tinybird.co
- Add your Tinybird workspace admin token in the Password field
- Save and test.
You're now ready to build dashboards and alerts on top of your Tinybird data sources and pipes.
Limitations & what's next
The ClickHouse HTTP interface currently supports read-only queries. You can query, visualize, and analyze your Tinybird Data Sources, but you can’t INSERT, UPDATE, or DELETE.
Other limitations that are worth noting are:
- System tables: Some advanced system tables (
query_log,parts_log, etc.) aren't fully supported yet, but are coming soon. - Compression: Not all ClickHouse compression modes are supported. We're working to improve this.
- Self-hosted infra: Support for self-hosted ClickHouse interface is also coming soon.
With the ClickHouse native HTTP interface, Tinybird now speaks the language of the many of the ClickHouse compatible tools you already use. This means simpler integrations, faster setup, and more options for how you query and visualize your data.
👉 Read the full docs and examples on how to use the ClickHouse HTTP Connector with Tinybird.
