I replaced Postgres and ClickHouse with one binary for web analytics
Single binary replaces Postgres, Redis, and ClickHouse stack for web analytics.

Single binary with zero runtime dependencies beats Docker-based competitors; tunnel architecture solves remote CH access elegantly.
ClickHouse operators, analytics engineers, and data teams needing local workspace + remote ClickHouse governance
DataGrip · QueryPie · DBeaver
I've been building CH-UI, an open-source ClickHouse workspace. Today I'm releasing v2, which is a complete rewrite from scratch.
v1 was a React SPA served via Docker. v2 is a single Go binary with an embedded Svelte frontend — download, run, open browser. No Docker, no Node, no runtime dependencies.
What it does:
- Multi-tab SQL editor with autocomplete - Database/table explorer with data preview - Saved queries - Tunnel architecture: run `ch-ui server` on one machine and `ch-ui connect` next to your ClickHouse — secure WebSocket tunnel, no VPN needed - Self-update (`ch-ui update`), OS service install (`ch-ui service install`)
Pro edition adds dashboards, scheduled jobs, AI assistant (BYO provider), governance (lineage, access matrix, policies), and alerting.
Core is Apache 2.0. Pro uses a signed license model (Ed25519).
Stack: Go 1.24, Svelte 5, SQLite (embedded state), Chi router, WebSocket tunnels.
GitHub: https://github.com/caioricciuti/ch-ui Install: `curl -fsSL https://ch-ui.com/install.sh | sh`
Happy to answer questions about the architecture, the tunnel model, or the open-core licensing approach.
Single binary replaces Postgres, Redis, and ClickHouse stack for web analytics.
Self-hosted Workspace with per-user SQLite and Yjs real-time collaboration.
Turns saved SQL queries into authenticated API endpoints instantly.
Single binary replaces Plausible/Fathom stack; embedded DuckDB kills infra tax.
GUI wrapper around imapsync for Gmail migrations when Google's own tool fails.
Sheets and charts in markdown is neat, but Obsidian and Notion already own this space.