useFrontendTool.mdx

title	useFrontendTool
description	The useFrontendTool hook allows the Copilot to execute tools in the frontend.

`useFrontendTool` is still supported, but we recommend migrating to [`useFrontendTool`](/reference/v2/hooks/useFrontendTool) from the v2 API, which uses Zod schemas for parameters.

useFrontendTool allows you to define executable actions that the AI can call with a handler function. This is the primary way to give your AI agent the ability to perform actions in your application—whether that's updating state, making API calls, or triggering side effects.

The hook requires three main pieces:

1. A name and description so the AI knows when to call it
2. A parameters definition describing what inputs the tool accepts
3. A handler function that executes when the AI calls the tool

Optionally, you can provide a render function to display custom UI showing the tool's execution status and results in the chat interface.

Usage

Simple Usage

import { useFrontendTool } from "@copilotkit/react-core";

useFrontendTool({
  name: "sayHello",
  description: "Say hello to someone.",
  parameters: [
    {
      name: "name",
      type: "string",
      description: "name of the person to greet",
      required: true,
    },
  ],
  handler: async ({ name }) => {
    alert(`Hello, ${name}!`);
  },
});

With Custom UI Rendering

import { useFrontendTool } from "@copilotkit/react-core";

useFrontendTool({
  name: "showWeatherCard",
  description: "Display weather information for a location",
  parameters: [
    {
      name: "location",
      type: "string",
      description: "The location to show weather for",
      required: true,
    },
    {
      name: "temperature",
      type: "number",
      description: "Temperature in celsius",
      required: true,
    },
  ],
  handler: async ({ location, temperature }) => {
    // Fetch and return weather data
    return { location, temperature, conditions: "Sunny" };
  },
  render: ({ args, status, result }) => {
    if (status === "inProgress") {
      return <div>Loading weather for {args.location}...</div>;
    }
    if (status === "complete" && result) {
      return (
        <WeatherCard
          location={result.location}
          temperature={result.temperature}
          conditions={result.conditions}
        />
      );
    }
    return null;
  },
});

Generative UI

This hook enables you to dynamically generate UI elements and render them in the copilot chat. For more information, check out the Generative UI page.

Migration from useCopilotAction

If you're migrating from useCopilotAction, here are the key differences:

1. The render component props include name and description

Migration Example

// Before with useCopilotAction
useCopilotAction({
  name: "addTodo",
  parameters: [
    {
      name: "text",
      type: "string",
      description: "The todo text",
      required: true,
    },
  ],
  handler: ({ text }) => {
    addTodo(text);
  },
});

// After with useFrontendTool
useFrontendTool({
  name: "addTodo",
  parameters: [
    {
      name: "text",
      type: "string",
      description: "The todo text",
      required: true,
    },
  ],
  handler: ({ text }) => {
    addTodo(text);
  },
});

Parameters

The name of the tool. A description of the tool. This is used to instruct the Copilot on how to use the tool. Array of parameter definitions for the tool. Each parameter object should have: - `name` (string): The parameter name - `type` (string): The parameter type (e.g., "string", "number", "boolean", "string[]", "object") - `description` (string): A description of what the parameter is for - `required` (boolean): Whether the parameter is required - `properties` (array, optional): For object types, define nested properties using the same schema

Simple example: [{ name: "query", type: "string", description: "The search query", required: true }]

Nested example:

[
  {
    name: "user",
    type: "object",
    description: "User information",
    required: true,
    properties: [
      { name: "name", type: "string", description: "User's name", required: true },
      { name: "age", type: "number", description: "User's age", required: false }
    ]
  }
]

The handler function that executes the tool logic. Whether to report the result of the tool call to the LLM which will then provide a follow-up response. Pass `false` to disable. A React component that renders custom UI for the tool. Whether the tool is available. Set to "disabled" to prevent the tool from being called.