// the find
caioricciuti/ch-ui
Use CH-UI to work with your data from Click House self-hosted with a user-friendly interface. CH-UI is a modern and feature-rich user interface for ClickHouse databases. It offers an intuitive platform for querying ClickHouse databases, executing queries, and visualizing metrics about your instance.
CH-UI is a self-contained ClickHouse management platform — SQL editor, dashboards, AI assistant, data pipelines, and dbt-style models — shipped as a single Go binary with an embedded Svelte 5 frontend. It targets developers and data engineers who run self-hosted ClickHouse and want a proper workspace rather than the bare-bones default UI. The free community tier is genuinely complete; Pro gates scheduling, governance lineage, and alerting.
Single-binary distribution with no external runtime dependencies is a real operational win — download, run, done. Streaming query results via SSE means long-running ClickHouse queries don't time out at the HTTP layer, which is a real pain point with naive approaches. The query cost estimation (estimated rows/parts before execution) mirrors BigQuery's dry-run UX and is genuinely useful for columnar DBs where bad queries are expensive. The WebSocket tunnel for remote ClickHouse access is architecturally clean — the agent only needs outbound connectivity, so it works behind NAT without firewall changes.
SQLite as the sole state store means multi-instance or HA deployments aren't possible without external coordination; if that SQLite file corrupts, you lose dashboards, saved queries, and Brain chat history. The dual-license split (Apache 2.0 community core vs BUSL 1.1 Pro features) is standard but the boundary isn't obvious from the README — scheduling and alerting being Pro-only will surprise teams that assumed 'open source' meant everything. Dashboard chart types are limited to uplot-backed options; anyone coming from Grafana will miss histogram panels, heatmaps, and table visualizations. At 680 stars and no sign of a maintained test suite in CI beyond unit tests, adoption is still narrow enough that you'd be betting on the author's continued interest.