Vercel launches unified reporting for AI Gateway usage across providers, users, and pricing tiers

What Vercel posted on X

On March 25, 2026, Vercel introduced a Reporting API for AI Gateway, highlighting the ability to track customer usage with tags, understand internal spend, compare pricing tiers, and calculate margins on AI features. The X post positioned the release as an operations layer for teams that already route inference through AI Gateway but still lack a single view of usage economics.

That matters because AI product costs are often fragmented across providers, credentials, and dashboards. Once teams support both managed credentials and bring-your-own-key traffic, it becomes much harder to answer basic questions like which feature is expensive, which customer is driving usage, or whether a pricing tier still works.

What the blog adds

Vercel says the Custom Reporting API is available in beta for Pro and Enterprise plans. The API gives programmatic access to cost, token usage, and request volume across AI Gateway traffic, including BYOK requests. According to the blog, teams can break down results by model, provider, user ID, custom tags, and credential type, which makes it easier to analyze spend per customer, per feature, or per pricing plan from one endpoint.

The post also explains how Vercel wants teams to use it. Developers can attach user and tags metadata through the AI SDK, Chat Completions API, Responses API, OpenResponses API, and Anthropic Messages API, and then query a unified reporting endpoint to retrieve the aggregated results. Vercel is trying to move reporting from spreadsheet-style reconciliation into the normal runtime and analytics flow of an AI product.

Vercel includes a business case too. The company says one AI platform serving more than 200,000 users replaced a separate proxy layer during the private beta and saved more than $80,000 while consolidating request management and cost tracking. That example is notable because it frames reporting not as back-office finance tooling, but as infrastructure that changes how teams route traffic and manage unit economics.

Why this matters

The broader signal is that AI infrastructure vendors are moving past simple model access and into margin management. Multi-model products need analytics that connect raw usage to actual features, customers, and pricing tiers rather than just per-provider bills.

If Vercel's reporting layer works as advertised, it gives teams a cleaner way to detect spend spikes, price AI features more intelligently, and decide when BYOK or routing changes make operational sense. That makes AI Gateway meaningfully more complete as a production platform, not just a routing abstraction.

Sources: Vercel X post · Vercel blog post