The real-time database for AI conference

Come check-in for the event and join us for coffee, breakfast and networking.

The ClickHouse Platform for AI is an open, end-to-end stack that gives agents the real-time analytics they demand through MCP, the CLI, and ClickHouse Agents powered by Claude. See how to use and build agents on ClickHouse in minutes, on the data you already have.

ClickStack is the turnkey, open-source observability stack built on ClickHouse and OpenTelemetry - full-fidelity, no sampling, no rollups, and MCP-native so users and agents can reason over your telemetry directly. See how it correlates logs, traces, and metrics in real time, at any scale.

Using ClickHouse Observability for the Australian Open, Winter Olympics, State of Origin and more.

The hardest part of an AI product isn't shipping it, it's knowing it's good, and keeping it that way. Langfuse turns production behavior into improvement material: trace, evaluate, improve, ship. Now part of ClickHouse, it pairs a ClickHouse-native V4 with code-based evals and CI/CD quality gates.

Four years of JOIN improvements, full read-and-write support for Iceberg and Delta Lake, and distributed query execution, all in one place. See how we've supercharged ClickHouse to analyse data wherever it lives, and query petabytes across your lake and warehouse at ClickHouse speed.

Buildkite’s Test Engine helps engineering teams make builds faster and more reliable with real-time test analytics and intelligent test splitting. Used by customers including OpenAI, Cursor, Slack, and Airbnb, the platform has ingested over a trillion test results into ClickHouse Cloud.

In this talk, we’ll explore the architecture behind that scale: ClickHouse Cloud on AWS operating alongside Rails, Flink, and PostgreSQL. We’ll trace how ClickHouse replaced workloads previously spread across PostgreSQL, Redis, DynamoDB, and Athena, and examine the design choices that made the transition practical. These include using ClickHouse’s JSON data type to split flexible key-value test tags into separate columns on disk, Materialized Views to accelerate important queries, and a soon-to-be-open-source Ruby ClickHouse query builder. Attendees will leave with practical ideas for introducing ClickHouse into an existing application stack and using test data to shorten build times.

The software and AI industry has landed on the “best-of-breed” data stack. Postgres powering transactional workloads and ClickHouse powering analytical workloads. Each doing what they are best suited for.

Neara builds physics-based digital twins that tell utilities which poles fail in a storm and which trees threaten the line. Ironically, we couldn't see inside our own platform: five tools, zero distributed tracing, and model builds that burst to a thousand machines. This is how we rebuilt it on ClickHouse and ClickStack — schema, collector, and all.

Real-time analytics. Observability. Data warehousing. LLM Observability. In this 3-hour workshop, you'll see ClickHouse power all of them — live. We start from the basics, demo each use case (including LLM tooling), then crack open the hood to understand the architecture that makes sub-second queries at petabyte scale possible. Whether you're evaluating ClickHouse Open Source or ClickHouse Cloud, you'll leave with a clear picture that ClickHouse is the right tool for you.

Enjoy light food and drinks, and hang out with the ClickHouse community!