Convex Cost

Before a Convex app ships, the expensive part is usually one or two functions quietly reading far more data than anyone expected, and convex-cost goes after exactly that. It pulls bytes and documents-read figures from the deployment's insights, multiplies them by call volume to rank which functions actually drive spend, and projects how each one scales as tables grow. A full-table `.collect()` is flagged as a cost that compounds with data, while an indexed `.take(n)` stays flat. For each driver it names the cheapest fix and hands off to convex-expert or convex-advisor for the change. It also forces an explicit yes on any metered action, like a plan change or domain purchase. It's read-only, so it's safe to run against prod. The one caveat: it gives curves rather than dollar totals, so you still check the pricing page for absolute numbers.

skills.sh installs: 41.1K
GitHub stars: 69
last updated: Oct 26, 2026

View on GitHub

Install

$ npx -y skills add convex/agent-skills --skill convex-cost --agent claude-code

Installs into .claude/skills of the current project.

Files

Preview what this app will cost

Cost surprises come from a handful of functions reading far more data than anyone realized — the same read-heavy patterns convex-advisor flags for perf, seen through the money lens. This capability makes spend legible: it reads the deployment's own bytes/documents-read evidence, attributes it to the functions driving it, projects how it grows with traffic, and names the cheapest fix. It also carries the confirm-cost discipline (Supabase's structural consent for paid actions): before anything metered, state the price and get an explicit yes.