Package version:

Interface Agent

Represents an agent that can call the model and use tools.

interface Agent {
    createdAt: Date;
    description: null | string;
    id: string;
    instructions: null | string;
    metadata: any;
    model: string;
    name: null | string;
    object: "assistant";
    responseFormat?: null | AgentsResponseFormatOption;
    temperature: null | number;
    toolResources: null | ToolResources;
    tools: {};
    topP: null | number;
}

Properties

createdAt: Date

The Unix timestamp, in seconds, representing when this object was created.

description: null | string

The description of the agent.

id: string

The identifier, which can be referenced in API endpoints.

instructions: null | string

The system instructions for the agent to use.

metadata: any

A set of up to 16 key/value pairs that can be attached to an object, used for storing additional information about that object in a structured format. Keys may be up to 64 characters in length and values may be up to 512 characters in length.

model: string

The ID of the model to use.

name: null | string

The name of the agent.

object

The object type, which is always assistant.

responseFormat?: null | AgentsResponseFormatOption

The response format of the tool calls used by this agent.

temperature: null | number

What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic.

toolResources: null | ToolResources

A set of resources that are used by the agent's tools. The resources are specific to the type of tool. For example, the code_interpreter tool requires a list of file IDs, while the file_search tool requires a list of vector store IDs.

tools: {}

The collection of tools enabled for the agent.

topP: null | number

An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered.

We generally recommend altering this or temperature but not both.