Chat
Talk to any model in the catalog through the same governed gateway your API keys use. This page is a local demo: nothing you type leaves your browser.
Pick a model above and say something. When the console ships, this box will talk to the real gateway, with your budgets, locks and cost tracing applied to every message.
Same gateway as your API
Chats run through the identical routing, retry and guardrail stack as production requests, so what you test is what you ship.
Switch models mid-thread
Change the model between messages and compare answers side by side without losing the conversation.
Every message costed
Each reply lands in the same ledger as your API calls, with the exact provider price attached.
Ready for the real thing?
Grab an API key and point any OpenAI-compatible SDK at the gateway; the chat console follows.