Experiential Labs

Experiential Labs runs an open source AI gateway that puts every model behind one API key and one endpoint. Hosted providers, your own provider keys and your own GPUs are all reached through the same compatible interface with streaming included, and routed tokens are billed at the provider's own list price with nothing added on top. The gateway itself can be self hosted, and the company earns on hosted inference and paid plans instead of a routing margin. It is aimed at teams running agents across several model providers who need routing, access control and attribution in one place. A console reports every request and every dollar by agent, by person, by model or by day, with budgets that stop requests at the cap, allowlists, live request logs and dashboards covering spend, latency, token volume, cache hit rate and active keys. Above the routing sits an intelligence layer: it watches traffic and flags when switching models would win, including models released that day, identifies caching opportunities and shows what they would save, and can train a smaller model on your own traffic that is then served through the same endpoint once you turn it on. Free costs $0 per month and includes 500 credits once a card is verified with a one time $1 charge. Paid tiers are determined by monthly spend: $20 to $199 per month is Pro, $200 to $1,999 is Max, and $2,000 and up is Ultra, with included credits running from 2,000 to 1,000,000 per month and each step raising rate limits and model discounts. Enterprise is custom and adds single sign on, advanced role based access control, private networking, data residency and committed credits. The company states it is backed by Y Combinator.