Ramp launches its own AI model router, called Router
Corporate expense platform Ramp launches Router, an API-based AI model routing service, free through 2026 with a $26 credit in the U.S.
Stock photo for illustration only, not from the actual event
- Ramp launches Router, an AI model routing service enabling API switching between LLMs
- Available exclusively in the U.S., free through 2026 with a $26 launch credit
- Features a dedicated dashboard tracking token spend, cost, latency, and fallback attempts
- Retains model inputs and outputs for one year with PII removal before product improvement
Corporate expense management platform Ramp is making its move into the AI inference toll house space, following closely behind Stripe. On Wednesday evening, the company officially launched its proprietary AI model routing service named Router, designed to let users and enterprises access and switch between various large language models via an API.
Ramp stated that the routing service has been built and battle-tested in-house for its own AI operational needs over the past three years. Functionally, Router operates in a manner quite similar to OpenRouter, although the latter currently supports a significantly broader selection of AI models.
Stock photo for illustration only, not from the actual event
The service is currently restricted to users in the United States. It is free to use for the remainder of 2026—though users are still responsible for underlying AI model inference costs—and comes bundled with a $26 launch credit offer. The company has not yet disclosed pricing details for next year.
Entering the model routing arena presents a strategic two-pronged opportunity for Ramp. Beyond tapping into the booming AI inference market, it allows the company to bundle a model routing solution directly with its existing product suite, which includes AI token usage monitoring and spend management. If Router succeeds in attracting developers much like OpenRouter has, it could establish long-term ties with AI labs and inference providers worldwide, serving as a powerful acquisition funnel for its core expense management products.
Administrators and developers are provided with a comprehensive dashboard that displays token spend, total cost, latency metrics, fallback attempts, and other granular operational details.
On the data governance front, Router implements an opt-out data retention policy. The platform records model inputs, outputs, and tool calls for a default period of one year, though Ramp emphasizes that personally identifiable information will be stripped out before any content is utilized for product enhancement.
Source: TechCrunch
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment