ClickHouse  BUILD Workshop

Build it,
end to end.

In one three-hour sitting, you take a real analytics app live on ClickHouse Cloud — hands-on, module by module, with your AI coding agent working alongside you.

3 hours · start to finish100% hands-on$0 on trial creditsbring your own agent
Your computer

Choose once. The workshop keeps the right setup visible on every page.

Not a lecture. You leave with a running prototype.

Most workshops hand you slides. This one hands you a real, working NYC-taxi ride-hailing app — React front end, FastAPI back end, a Postgres source — and, module by module, you move its data and its intelligence onto ClickHouse Cloud.

You never build from scratch and you never fall behind. Every command is copy-paste and your coding agent handles the typing, so you can focus on learning each piece. Each module states exactly what you need in place — so if you get stuck, you catch up in seconds.

When the session ends, the trial keeps running, everything you built keeps running on it, and the repository is yours to take back to your team and demo live.

local applive on ClickHouse Cloud
static datareal-time CDC stream
SQL by handconversational BI
blind incidentsAI SRE + traced chat

What you walk away with

Everything you'll have running by the end.

All on your own ClickHouse Cloud trial — you connect it, module by module, during the session.

01

A live ops dashboard

Over 3.2M+ real NYC taxi rows, seeded straight from object storage — and fast enough to feel it.

02

A real-time CDC pipeline

Your own managed Postgres streaming continuously into ClickHouse through a ClickPipe.

03

Conversational BI

Ask your data questions in plain language with ClickHouse Agents — governed, no SQL required.

04

Full observability + an AI SRE

App traces and logs flow into ClickStack / HyperDX; your agent builds a dashboard and alert over them, then diagnoses a live incident.

05

Traced AI chat

An in-app AI chat with every turn, generation, and cost traced end to end in Langfuse.

+

The whole repo

Yours to keep, extend, and point at your own data after the room clears.

One platform, end to end

Every ClickHouse product you'll use.

A guided tour across ClickHouse Cloud — you connect and configure each piece yourself, from ingestion to AI.

Platform + database
ClickHouse Cloud
The managed service everything runs on — millions of rows, sub-second queries.
Ingestion
ClickPipes
Postgres CDC that streams your live rows into ClickHouse.
Source database
Managed Postgres
Postgres, managed by ClickHouse — your CDC source, created with clickhousectl.
Conversational BI
ClickHouse Agents
Ask your data questions in plain language, governed by RBAC.
Observability
ClickStack + HyperDX
Your app's traces and logs, searchable in the HyperDX UI.
LLM observability
Langfuse
Trace every AI chat turn — generations, latency, and cost — end to end.
Agent integration
ClickHouse MCP
The MCP endpoints your coding agent connects to, over OAuth, to drive it all.

Plus the one non-ClickHouse service the app calls: OpenAI (the chat model behind the in-app AI chat).

The route · ~2h30 hands-on

Ten stops, in order.

A three-hour session: ~2h30 of building, plus the opening, transitions, and the live finale demos.

  1. 00

    Setup

    25 min

    Accounts, keys, tools, agent skills, and the app running locally — all wired up and verified.

  2. 01

    ClickHouse Cloud

    15 min

    Create the schema, seed 3.2M+ historical rows from object storage, and feel the query speed.

  3. 02

    Base app

    5 min

    Tour the running app now that it has data: the Ops and Historical dashboards and the chat panel.

  4. 03

    Real-time CDC

    20 min

    Stream live rows from your own managed Postgres into ClickHouse with a Postgres CDC ClickPipe.

  5. 04

    ClickHouse Agents

    10 min

    Conversational BI: create an agent over your taxi data and explore it in natural language.

  6. 05

    ClickStack

    15 min

    Enable ClickStack and send your app's traces and logs to HyperDX.

  7. 06

    AI SRE

    15 min

    Connect your agent to the ClickStack MCP and have it build an SRE dashboard and an alert.

  8. 07

    Break and fix

    20 min

    Inject a realistic fault, diagnose it from telemetry with an AI SRE, and ship the fix.

  9. 08

    Chat and Langfuse

    15 min

    Use the in-app AI chat and follow its traces, generations, and costs in Langfuse.

  10. 09

    Wrap-up

    10 min

    Review what you built, take it home, and extend it to your own data.

Runs with an instructor or fully self-paced. The primary path is self-serve — your coding agent doubles as your instructor, and a troubleshooting reference covers every failure seen in testing.

Before you join

Who it's for, and what to bring.

Who it's for audience

  • Solutions architects and pre-sales engineers at partners
  • Data and platform engineers evaluating ClickHouse
  • Technical teams at partners' customers who want the full stack, hands-on
  • Comfortable in a terminal and working alongside an AI coding agent

What to bring prerequisites

  • A laptop with Docker running
  • Your own agentic coding tool — Claude Code, Cursor, Codex CLI, or Windsurf — signed in on an active plan
  • A ClickHouse Cloud account with trial credits (created in prework)
  • An LLM for the chat module — bring your own OpenAI-compatible model (Azure, OpenRouter, vLLM, or local), or ~$5 of OpenAI credit that stretches across many runs

How it runs format

  • 100% hands-on, module by module — you configure and connect, you don't edit app code
  • Every command and query is copy-paste
  • Never stranded: each module names a checkpoint you can start fresh from
  • Your dashboard is live on real data inside the first 40 minutes

Ready to build?

Bring your team, your laptops, and your coding agents. Leave with a real-time, observable, AI-assisted analytics app on ClickHouse Cloud.

$0ClickHouse Cloud trial credits cover the whole workshop.
30 daysYour trial and everything you built keep running afterward.
YoursThe repository is yours to keep, extend, and demo.