@squidcloud/client
    Preparing search index...

    Interface AiApiOptions

    Options for configuring AI API call execution.

    interface AiApiOptions {
        instructions?: string;
        maxIterations?: number;
        memoryOptions?: AiAgentMemoryOptions;
        stageTimeoutMs?: number;
    }
    Index
    instructions? maxIterations? memoryOptions? stageTimeoutMs?
    instructions?: string

    Custom instructions appended to the endpoint-selection, body-construction, and response-analysis prompts.

    maxIterations?: number

    Maximum number of distinct API calls the AI is allowed to make for one user prompt. Default 4.

    memoryOptions?: AiAgentMemoryOptions

    Conversation memory configuration. When provided, prior user prompts and AI answers under the same memoryId are provided as additional context to the AI, enabling cross-prompt follow-ups Defaults to no memory.

    stageTimeoutMs?: number

    How long, in milliseconds, a single AI call of the loop may wait for the provider — endpoint selection, request-body creation and response analysis each get this budget on their own, so the whole call may run for a multiple of it. Values below the chat minimum of 20 000 ms are raised to it. Defaults to the standard chat timeout, or to the timeout of the chat this call runs inside when the caller is a nested one.

    Settings
    Member Visibility

    On This Page
    Properties