Show HN: Marginal – See which customers and features drive your AI API costs
Marginal is a new tool that lets companies attribute AI API costs to specific customers and product features in real time.
Marginal launched publicly on August 19, 2026, via a Show HN post, offering a cost-attribution dashboard for companies spending on AI APIs.
The tool is designed to answer a question common among AI-powered SaaS businesses: which customers or product features are consuming the most tokens — and therefore driving the most cost. Per the Show HN, the product includes an Overview, an Insights view, and an Explorer, and accepts real usage events sent from a company's own systems. The author, posting under the handle jithinlalk25, shared a demo login backed by a seeded fictional-company workspace so prospective users can explore the interface before signing up.
Marginal sits at the intersection of AI observability and cost management, a niche that has grown as token spend on models from OpenAI, Anthropic, and others has become a meaningful line item for product teams. By tagging API calls with customer or feature identifiers, the tool aims to give engineering and finance teams a breakdown that raw provider dashboards don't offer.
No funding, team size, or founding date was disclosed in the Show HN post, and no additional sources have confirmed further details. The product appears to be at an early, self-serve stage, with sign-up open at marginalhq.com.
No comments yet — start the thread.