Corporate expense management company Ramp has introduced a new AI model routing service called Router, which allows users and businesses to access and switch between various large language models (LLMs) through a single API. The company has been using this routing technology in-house for three years before launching it as a public product, currently available only in the U.S. Router is free to use through the end of 2026, with users responsible for paying inference costs, and includes a $26 credit at launch. Pricing for 2027 has not yet been disclosed.

Router competes with services like OpenRouter by providing API-driven access to AI models from providers such as OpenAI, Anthropic, DeepSeek, Moonshot, Minimax, Nvidia, xAI, and Z.ai. It offers customers customizable “strategies” that route queries based on preferences for cost, usage flexibility, benchmarks, or complexity of the tasks. Users also gain visibility through a dashboard that tracks token spending, latency, fallback attempts, and other usage metrics. Importantly, Router implements a data retention policy where inputs and outputs are stored for one year by default but removes personally identifiable information before using data to enhance the service.

Ramp sees an opportunity to capitalize on the growing AI inference market while integrating Router with its existing financial oversight products, which monitor AI token usage and spending. With the service, Ramp expects to strengthen ties with AI labs and inference providers and provide its traditional corporate users a seamless way to manage AI model usage and expenses. This expansion complements Ramp's broader fintech role, underscored by the company’s recent $750 million funding round valuing it at $44 billion.

By offering Router, Ramp positions itself to compete in the emerging AI expense management sector alongside companies like Stripe and Databricks, which are also developing AI gateways and expense tracking tools. This initiative reflects a wider trend of fintech firms leveraging AI-related infrastructure to deepen customer engagement and potentially unlock new revenue streams linked to enterprise AI consumption. Router’s ability to balance cost efficiency and quality through model routing could prove attractive as businesses increasingly rely on AI services.