About Xantly

Xantly Inc. builds AI gateway infrastructure. Our product is a single OpenAI-compatible API endpoint that reaches 10,000+ language models, routes each request to the model that fits it, caches what it can, remembers what it should, and fails over when a provider does not answer.

Why we build it

Teams shipping AI features end up maintaining an integration per provider, a fallback path per integration, and a spreadsheet of which model to use for what. That work is undifferentiated: it is the same problem at every company, it has to be redone every time a provider ships a new model, and it quietly grows into the most expensive line on the infrastructure bill.

Xantly makes that a configuration change. You point base_url at https://api.xantly.com/v1, keep your existing SDK, and the routing, caching, memory, failover and cost accounting come with it. Customers typically see 60 to 80% lower spend, which comes from three stacked mechanisms: routing to the cheapest model that clears the quality bar, serving repeated prompts from cache at a flat $0.25 per million tokens, and not paying twice for requests a provider failed.

How we charge

One 3% platform fee at top-up, and usage billed at exactly what the upstream provider charges. No subscription, no seats, no model gates. We are paid when you add credits, not by marking up your tokens, so routing you to a cheaper model does not cost us anything. Full detail is on the pricing page.

Where we are

Xantly Inc. is a company registered in the United States, and the platform is operated from United States infrastructure. Requests can be routed across several clouds and regions, and enterprise agreements can constrain which regions your traffic runs in. Our sub-processors, data flows and retention periods are enumerated in the Privacy Policy and the Data Processing Agreement.

Security and compliance

We publish our Acceptable Use Policy, Terms of Service and Cookie Policy. Privileged actions are recorded in a hash-chained, tamper-evident audit trail. Single sign-on over OIDC with SCIM provisioning is self-serve for org admins. Report a vulnerability to security@xantly.com.

Talk to us

General enquiries go to hello@xantly.com. The full list of addresses, including privacy, legal, billing and security, is on the contact page.