vercel metrics pulls your Web Analytics data straight from the command line. No dashboard tab-switching, no API wrangling — page views, visitors, and custom events come back as CLI output you can pipe, parse, or hand to an agent.
The command supports querying across dimensions, filters, and date ranges. Vercel’s changelog entry frames the feature around three use cases:
- Traffic analysis — which pages gained the most views this week
- Campaign attribution — which UTM campaigns drove the most signups this month
- Device segmentation — compare conversion events between mobile and desktop users
The real hook isn’t the CLI itself — it’s what happens when you give a coding agent access to it. Vercel explicitly positions vercel metrics as a tool an AI agent can call to answer performance questions autonomously. Instead of a developer manually pulling reports, an agent running inside Cursor or Claude Code can execute vercel metrics, interpret the output, and fold the insight into whatever task it’s already working on. Debugging a drop in signups? The agent checks UTM performance and cross-references with recent deploys without you alt-tabbing to an analytics dashboard.
This fits a pattern across the Vercel platform: making infrastructure queryable by machines, not just humans. The vercel CLI already handles deploys, environment variables, and domain config. Adding analytics closes a gap — agents can now observe production behavior, not just ship code to it.
TIP
Point your coding agent’s tool-use config at the Vercel CLI and let it run vercel metrics queries mid-task. The full docs list every supported metric, dimension, and filter option.
Available now in the Vercel CLI. You need Web Analytics enabled on your project — the command queries the same data you’d see in the Vercel dashboard’s Analytics tab.