Free playground

Send raw input and inspect the intent identified by a real model.

IntentParse funds a low-cost model for this stateless demonstration, limited to one inference per network origin every ten minutes. Production API inference remains customer-paid BYOK.

Schema routing

The router selects the domain before running the specialized schema.

Raw user message56 / 1,000

Free playground inference
IntentParse funds a low-cost LLM for this demonstration. Its credential stays server-side; during testing, each network origin may run one inference every 30 seconds.

Retention: stateless parse. The input and parsed state are not retained.

1 · Parse2 · Persistence off3 · Grounding —

Typed state will appear here.