Ramp, the corporate spend management company, has launched Router, a service that lets businesses send queries to large language models from multiple providers through a single API and switch between them at will, TechCrunch reports. At launch, Router supports eight providers: OpenAI, Anthropic, DeepSeek, Moonshot, Minimax, Nvidia, xAI, and Z.ai.
The product is organized around routing strategies. Companies can steer traffic toward providers' cheaper flex-usage tiers, have Router pick models based on up to three benchmarks they select, or reserve expensive frontier models for only the hardest problems while cheaper models handle routine requests. Because everything flows through one API, teams can test a new model without rebuilding their integration. A dashboard tracks token spend, cost, latency, and fallback attempts — extending the spend-visibility approach of Ramp's core business into AI usage.
Router is free through the end of 2026, with customers paying providers' inference costs directly, and comes with a $26 launch credit. Pricing for 2027 has not been announced, and availability is currently limited to the United States. By default, the service retains model inputs, outputs, and tool calls for one year, and Ramp says it strips personally identifiable information before using that content to improve the product.
The launch is less of a leap than it might appear: Ramp has used the routing technology internally for three years, and the company positions Router alongside its existing AI token-monitoring and spend-management tools, giving finance teams one place to watch what their organizations spend on AI.
Founded in 2019 and headquartered in New York, Ramp built its business on corporate cards and expense automation before expanding into a broader finance platform. The company raised $750 million at a $44 billion valuation in June, and Router marks its most direct step yet into AI infrastructure — betting that the company already watching a business's spending is well placed to route its AI traffic too.