Metriq vs
Hex, Deepnote, Julius.

The most mature product in the category. Five years of polish on reactive DAG execution: cells re-run automatically when upstream changes, with proper dependency tracking. They also shipped agent memory in April 2026, closing one of the gaps we previously highlighted. Well-funded and execute well. If you need a polished, hosted, opinionated analytics product and your team doesn't care about lock-in, Hex is the strongest choice in the closed-source tier.

Strong real-time collaboration. Best-in-class block library for non-coding analysts. Their decision to open-source the notebook format in October 2025 was a meaningful step toward reducing lock-in. If your team is half analysts, half non-technical, Deepnote's block-based authoring is more accessible than ours.

Conversational analytics for non-technical users. If you don't want to see code at all, ever, Julius is the right product. It hides everything. We don't.