Try the lab
A private CHouse UI with sample data in your own GitHub Codespace: what it costs, what's inside, how to turn on AI and how to delete it.
The lab is a private copy of CHouse UI that runs in your own GitHub Codespace. It comes with a ClickHouse server, sample data and example content in every area, so you can try everything as admin before you install anything.
Open the lab in GitHub Codespaces
What you need, and what it costs
- A GitHub account. The codespace runs in your account. There is no hosted demo and no sign-up with us.
- About 3 minutes for the first start, while the codespace pulls the images and loads the data.
- Some of your free Codespaces quota. Personal GitHub accounts include free monthly compute and storage. The lab uses the 2-core machine, which spends 2 core-hours of that quota per hour it runs. GitHub doesn't charge you past your spending limit, which is $0 unless you raised it. See GitHub's Codespaces billing page for the current allowance.
- Optionally, your own AI key. It is needed only for the AI features (see below), and billed to you by your provider.
Only you can open your lab. Its forwarded address needs your GitHub sign-in, and its admin password and secrets are generated for that one codespace.
What's inside
Each codespace runs the published CHouse UI image, one ClickHouse server, a light query load and a one-off loader. The loader fills:
| Area | What you'll find |
|---|---|
| SQL workspace | shop and analytics databases with 2 million orders, a materialized view, a small finance database, and four shared saved queries. |
| Monitoring | A steady query load, new orders every 30 seconds, one slow query every 2 minutes and one failing query every 10 minutes. |
| Dashboards | Shop sales, and the ClickHouse overview, Query latency and Disk and tables templates. |
| Data | Two data-health promises. One holds; Shop orders are in EUR breaks on purpose, so there is an incident to investigate. |
| Alerts and schedules | A fleet threshold rule and a daily report that has already run once. |
| Access control | Teammates maya (analyst), devon (developer) and vic (viewer), and a data access policy that limits analysts to the sample databases. |
| Context | Descriptions, owners and a revenue metric for the sample tables. |
The in-app Getting Started guide opens on first sign-in and walks through
every area. The codespace also opens a short guide, WELCOME.md, with things
to try.
Sign in
When the lab is ready, the terminal prints the address and a password made for that lab:
CHouse UI lab is ready.
Open: https://<your-codespace>-5521.app.github.dev
Sign in: admin / Lab-…-Ch1!
teammates maya (analyst), devon (developer) and vic (viewer) use the same passwordThe browser tab opens on its own. If a popup blocker stops it, open the
Ports panel and click the globe icon next to CHouse UI (5521). The
sign-in is also saved in .devcontainer/lab/.state/login.txt.
Turn on AI
The AI SRE, Optimize with AI, Chobi and chat need a model. The lab uses your own key:
- Add a Codespaces secret named
ANTHROPIC_API_KEYorOPENAI_API_KEYin your Codespaces settings, with access todaun-gatal/chouse-ui. If you already have one with that name, the lab finds it. - Create the codespace, or rebuild it (<kbd>F1</kbd> › Codespaces: Rebuild Container).
The lab registers the key as the default model: Claude Haiku for Anthropic, or GPT-5 mini for OpenAI. Anthropic is used when both are set. You can change the model under Admin › AI models.
When you're done
A codespace stops after 30 minutes without activity. That is a setting in your GitHub account, not part of the lab. A stopped codespace still counts against your storage until it is deleted. Delete it at github.com/codespaces, or let GitHub remove it after your retention period (30 days by default). Deleting it removes everything, including any key you added to the lab.
Run the lab without Codespaces
The lab is a Docker Compose stack, so it also runs on any machine with Docker:
git clone https://github.com/daun-gatal/chouse-ui.git
cd chouse-ui
bash .devcontainer/lab/up.shOpen http://localhost:5521. If you open it from another machine, set
LAB_URL to the address your browser uses (for example
LAB_URL=http://my-vm:5521 bash .devcontainer/lab/up.sh). Remove it with
docker compose -p chouse-lab down -v.
The lab is for trying CHouse UI, not for running it. To install it for your team, start with the quick start.