Voice Model-Ready Endpoint Console
Try a voice-ready endpoint and inspect the focused response that a voice model receives. Requests use your own API key and count against your rate limits.
The presets use London locations. Review your account's region and limits before changing them.
Use the key from your account. Submit your project requirements to discuss API access; eligible confirmed accounts can test London examples during review.
For voice models and voice agents, this endpoint reduces response tokens and keeps spoken-answer tool calls focused. Return up to the first two structurally unique route alternatives from /v1/routes in a normalized compact response for clients that need a smaller payload. Alternatives with the same ordered transit services, directions and boarding/alighting stops are deduplicated even when their departure and arrival times differ. Field names are declared once in *Fields arrays; data rows contain values in that order and connect through integer *Ref indexes. Geometry, intermediate stop objects and presentation-only metadata are omitted, while journey instructions, stop and route identifiers, directions, times, accessibility and alerts are retained for follow-up requests.
Request
Response
Response headers