Vloop API
    Preparing search index...

    Interface ChatStartOptions

    interface ChatStartOptions {
        cacheWarmerPingIntervalMs?: number | false | "auto";
        cwd?: string;
        driver?: string;
        driverOptions?: Record<string, string>;
        effort?: string;
        env?: Record<string, string>;
        gitRepoCheck?:
            | "auto"
            | "on"
            | "off"
            | "autoInside"
            | "autoInsideEvenIgnored";
        input?: ChatInputPart[];
        labels?: string[];
        model?: string;
        prompt?: string;
        promptFile?: string;
        resumeSessionId?: string;
        runner?: AgentRunner;
        serverSharing?: "shared"
        | "per-chat";
        timeoutMs?: number;
        title?: string;
    }
    Index
    cacheWarmerPingIntervalMs? cwd? driver? driverOptions? effort? env? gitRepoCheck? input? labels? model? prompt? promptFile? resumeSessionId? runner? serverSharing? timeoutMs? title?
    cacheWarmerPingIntervalMs?: number | false | "auto"

    Cache warmer: "auto" (the default) warms before the cache retention ends, a number of ms warms that long after each cache use, and 0 or false turns it off.

    cwd?: string

    Working directory, relative to the project unless absolute.

    driver?: string

    Driver registered by the runner. Defaults to auto.

    driverOptions?: Record<string, string>

    Driver-specific settings.

    effort?: string

    Requested reasoning level.

    env?: Record<string, string>

    Environment variables for the AI process.

    gitRepoCheck?: "auto" | "on" | "off" | "autoInside" | "autoInsideEvenIgnored"

    Codex trusted-directory check policy. Defaults to auto; other runners ignore it.

    input?: ChatInputPart[]

    The parts of the first prompt, when it has images.

    labels?: string[]

    Workbench labels.

    model?: string

    Model understood by the chosen runner.

    prompt?: string

    Inline prompt. Used before promptFile when both are present. One of prompt or promptFile is required.

    promptFile?: string

    Prompt file to read when prompt is not supplied.

    resumeSessionId?: string

    Raw provider session ID to continue. Prefer ai.chatFromVloopexId() for chats managed by Vloopex.

    runner?: AgentRunner

    Runner to launch. Defaults to claude.

    serverSharing?: "shared" | "per-chat"

    Use shared for a provider server shared with other chats, or per-chat for a separate server. Defaults to shared.

    timeoutMs?: number

    Turn timeout in milliseconds.

    title?: string

    Workbench title. Defaults to the first prompt line.

    Settings
    Member Visibility

    On This Page
    Properties