Voice

Dry-run one case, saved or ad-hoc

SPENDS NOTHING. One dry run, from a saved case (test_case_id) or an ad-hoc input (input_message), never both.

Returns the run row with the composed preview: the instruction stack the agent would receive, the tools it could reach for and the grounding it could cite, judged against the assertion. No model call, no call placed, no message sent. triggered_by defaults to the key's identity.

post/voice/agents/{agent_id}/test

Request

  • Base URL: https://be.graph8.com/api/v1
  • URL: https://be.graph8.com/api/v1/voice/agents/{agent_id}/test
  • Auth: HTTP bearer

Path parameters

agent_idstring required

Voice agent id.

Headers

Authorizationstring nullable
X-Target-Org-Idstring nullable

Request body

mode'preview' | 'dry_run' nullable

preview | dry_run - both resolve configuration only.

test_case_idstring nullable
input_messagestring nullable
input_contextobject nullable
channel'chat' | 'inbound' | 'outbound' nullable
expected_outputunknown
assertion_type'exact' | 'contains' | 'not_contains' | 'expects_tool' | 'refuses' | 'cites_collection' | 'llm_judge' nullable
assertion_configobject nullable
triggered_bystring nullable

Response

Successful Response

dataVoicePayload required

One voice object - a collection, an agent, a route, a chat.

Open-shaped: voice adds fields to these bodies on its own release cadence and a dropped-or-added key must not be a 500 in this service.

Changes

No recorded changes to this operation across all 2 revisions of this API.