Corporate spend management platform Ramp has followed Stripe's lead and is now building a “toll booth” for AI inference.

On Wednesday evening, Ramp launched its own AI model routing service with a straightforward name: Router. Users and companies can use an API to switch between multiple large language models. Ramp says it has been using this router internally for the past three years.


Router is currently available only in the United States. It will be free for the remainder of 2026, with users only paying for model inference costs. Ramp is also offering a $26 startup credit. It hasn't disclosed pricing for next year.

Router works similarly to OpenRouter, though OpenRouter supports far more models than Ramp currently offers.

Router currently offers models from OpenAI, Anthropic, DeepSeek, Moonshot, Minimax, Nvidia, xAI, and Z.ai. It also provides several “strategies” that let customers route requests according to their preferences. For example, it can prioritize a provider's burstable usage tier, or let Router choose a model based on up to three user-specified benchmarks. It can also route only difficult tasks to pricier models, or make it easy to test models without switching.

Users get a dashboard to view token spend, costs, latency, fallback attempts, and other details.

One thing worth noting: Router retains data by default. Model inputs, outputs, and tool calls are logged for one year by default. Ramp says it will remove “personally identifiable information” before using this content to improve its products.

For Ramp, entering the model routing market presents a twofold opportunity: it can break into the hot AI inference market while giving existing customers a routing service that pairs perfectly with its own products—AI token usage monitoring and token spend management.

Additionally, if Router can become an attractive model testing ground like OpenRouter, Ramp could build long-term relationships with AI labs and inference providers around the world. That could help Ramp—which just raised $750 million at a $44 billion valuation in June—acquire new customers and find new sales entry points for its expense management products.