Sends a prompt to a specific miniapp generator deployment and returns a streaming response using Server-Sent Events. The response is a continuous stream of Server-Sent Events, not a single JSON payload. Each event contains a JSON object with type, message, and other fields specific to the message type. Requires authentication via API key in the request header.
@anthropic-ai/claude-code except ErrorMessage):
'assistant''user''result' and subtypes:
'success''error_max_turns''error_during_execution''system' and subtype 'init''error'API key to authorize requests
Prompt string to send to the deployment
Deployment ID (UUID). Required if name not provided.
Farcaster ID of the user; if not provided, namespace must be provided
x >= 03
Kubernetes deployment name. Required if deployment_id not provided.
Optional Kubernetes namespace. If not provided, will query for the active namespace for the given FID.
Optional conversation ID to continue an existing chat. If not provided, a new conversation will be created.
Optional Claude SDK session ID to resume an existing Claude Code session. Enables session-based conversation continuity.
System prompt variant to use. Defaults to stable if not provided.
canary, beta, stable Success