functai for TypeScript - v0.1.0
    Preparing search index...

    Interface ModuleSpec<I, O, A>

    What module(name, spec, run) takes besides its code: its interface, what it uses, and settings.

    interface ModuleSpec<
        I extends Fields,
        O extends Fields | undefined,
        A extends FieldSpec | undefined,
    > {
        adapter?: unknown;
        answerFrom?: AnyAIFunction | null;
        apiRetries?: number;
        approve?: ApproveSetting | null;
        cacheReplies?: CacheSetting;
        caller?: Record<string, unknown> | null;
        capabilities?: Capabilities | null;
        config?: Partial<Config> | null;
        definedIn?: string;
        description?: string;
        escalateBelow?: number | null;
        escalateTo?: string | object | null;
        includeFnName?: boolean;
        input: I;
        journal?: Journal | null;
        lm?: string | null;
        logCalls?: string | boolean | null;
        logContent?: LogContent | null;
        maxSteps?: number;
        maxTokens?: number | null;
        module?: "predict" | "cot" | null;
        observers?: readonly Observer[] | null;
        output?: A;
        outputs?: O;
        plugins?: readonly Plugin[] | null;
        programObservers?: boolean | null;
        programPlugins?: boolean | null;
        replicate?: number | null;
        retries?: number;
        router?: Router | null;
        seed?: number | null;
        stop?: readonly string[] | null;
        temperature?: number | null;
        template?: readonly Record<string, unknown>[] | null;
        toolErrors?: "report" | "raise";
        topP?: number | null;
        uses?: readonly (AnyAIFunction | AnyModule)[];
    }

    Type Parameters

    • I extends Fields
    • O extends Fields | undefined
    • A extends FieldSpec | undefined

    Hierarchy (View Summary)

    Index
    adapter? answerFrom? apiRetries? approve? cacheReplies? caller? capabilities? config? definedIn? description? escalateBelow? escalateTo? includeFnName? input journal? lm? logCalls? logContent? maxSteps? maxTokens? module? observers? output? outputs? plugins? programObservers? programPlugins? replicate? retries? router? seed? stop? temperature? template? toolErrors? topP? uses?
    adapter?: unknown

    The layout: "xml" (default), "chat", "json", an lmcc adapter, or an artifact.

    answerFrom?: AnyAIFunction | null

    The AI function whose answer, as it is written, is the module's answer (streaming.md, Views): a served module's caller sees that function's answer text as the module's, live.

    apiRetries?: number

    Re-sends after a transient provider error (default 3).

    approve?: ApproveSetting | null

    Who answers a tool call before it runs (contract/tools.md): a function asked at once, or a rule a person answers later ("changes", "all", a list of tool names or approval paths). No approval by default.

    cacheReplies?: CacheSetting

    Answer an identical request with the reply it got before (true: in this process's memory; or a store: a Map, Redis, …). Off by default. It returns identical samples too: leave it off where you want different ones.

    caller?: Record<string, unknown> | null

    Who is calling, added to FUNCTAI_CALLER: { kind: "agent", conversation: "…" }.

    capabilities?: Capabilities | null
    config?: Partial<Config> | null

    Any other lm15 Config fields.

    definedIn?: string

    The code module the call log files it under (program.module; default: the file's name, without extension).

    description?: string

    What it does, in words.

    escalateBelow?: number | null
    escalateTo?: string | object | null

    When the first model is less sure of its answer than escalateBelow (default 0.9), another answers instead: a model, or an AI function. The first needs to measure its confidence (a baked model, TypeSafe's Jev, config: { probabilities: "required" }).

    includeFnName?: boolean
    input: I

    Its inputs, by name: { message: t.string() }, zod, or any Standard Schema. {} for none.

    journal?: Journal | null

    Where each call tree's kept log is written while it runs: a store (best effort: the call never waits), or { store, mode: "required" } (the call waits until its events are kept, before its code runs, before each tool and before it returns; at most timeout ms, 30 s by default). null: none. A program's own setting cannot replace or remove a host's journal (JournalError journal-policy).

    lm?: string | null

    The model: "gpt-4.1-mini", "claude-haiku-4-5", "groq:openai/gpt-oss-120b", …

    logCalls?: string | boolean | null

    The call log: a folder, true (the default folder), or false. Unset: FUNCTAI_LOG_CALLS decides.

    logContent?: LogContent | null

    Which values the call log and the kept form of events keep: true, false (sizes, times and tokens only), or by field ({ transcript: false }; { "*": false, question: true } keeps only the question). It only ever removes: a value is written only when no layer drops it, and FUNCTAI_LOG_CONTENT=0 drops every value.

    maxSteps?: number

    Model calls per tool loop (default 8).

    maxTokens?: number | null
    module?: "predict" | "cot" | null

    "predict" (default) or "cot": reasoning before the answer.

    observers?: readonly Observer[] | null

    Receivers of the kept form of every event of the calls in scope (a page, telemetry). They add up over the layers (configure's replaces configure's: keep your own list to add one). A function is called soon after each event, off the call's own turn, from a bounded queue (an observer 10,000 events behind loses events); it still runs on this thread, so heavy work belongs in a Worker: an object with postMessage is posted each event. One that throws is warned about once and given no more events.

    output?: A

    Its one output (named result), what run returns.

    outputs?: O

    Several outputs, in order (the last is the answer): run returns a record of each by name.

    plugins?: readonly Plugin[] | null

    Plugins around the calls in scope (contract/plugins.md): they add up over the layers, the program's own first, the host's last.

    programObservers?: boolean | null

    false (a host's block or configure): the observers a program sets for itself are given no event; the host's still are. It only removes: a program's own true does not undo it.

    programPlugins?: boolean | null

    false (a host's layer): the program's own plugins do not run.

    replicate?: number | null

    The n-th independent answer to the same request (default 0): part of the reply cache's key, and nothing else.

    retries?: number

    Re-asks after an unreadable reply (default 1).

    router?: Router | null

    The lm15 router calls go through (keys, base URLs, a fake for tests).

    seed?: number | null
    stop?: readonly string[] | null
    temperature?: number | null
    template?: readonly Record<string, unknown>[] | null

    A chat template of lmcc messages, instead of a named layout.

    toolErrors?: "report" | "raise"

    "report" (the model sees a tool's error) or "raise".

    topP?: number | null
    uses?: readonly (AnyAIFunction | AnyModule)[]

    The AI functions and modules it calls: its version changes when they change.

    Settings
    Member Visibility

    On This Page
    Properties