Structured Output

The final building block we'll look at is structured outputs.

There are two ways to achieve structured output with the AI SDK.

Experimental output with generateText (this allows you to use tools with structured outputs)
The generateObject function

import { openai } from '@ai-sdk/openai'
import { generateText, Output, tool } from 'ai'
import 'dotenv/config'
import { z } from 'zod'
 
const main = async () => {
  const result = await generateText({
    model: openai('gpt-4o'),
    prompt: 'Get the weather in SF and NY, then add them together.',
    maxSteps: 3,
    tools: {
      addNumbers: tool({
        description: 'Add two numbers together',
        parameters: z.object({
          num1: z.number(),
          num2: z.number(),
        }),
        execute: async ({ num1, num2 }) => {
          return num1 + num2
        },
      }),
      getWeather: tool({
        description: 'Get the current weather at a location',
        parameters: z.object({
          latitude: z.number(),
          longitude: z.number(),
          city: z.string(),
        }),
        execute: async ({ latitude, longitude, city }) => {
          const response = await fetch(
            `https://api.open-meteo.com/v1/forecast?latitude=${latitude}&longitude=${longitude}&current=temperature_2m,weathercode,relativehumidity_2m&timezone=auto`
          )
 
          const weatherData = await response.json()
          return {
            temperature: weatherData.current.temperature_2m,
            weatherCode: weatherData.current.weathercode,
            humidity: weatherData.current.relativehumidity_2m,
            city,
          }
        },
      }),
    },
    experimental_output: Output.object({
      schema: z.object({ sum: z.string() }),
    }),
  })
  console.log(result.steps.length)
  console.log(result.experimental_output)
}
 
main()

The other option is the generateObject function:

import { openai } from '@ai-sdk/openai'
import { generateObject } from 'ai'
import 'dotenv/config'
import { z } from 'zod'
 
const main = async () => {
  const result = await generateObject({
    model: openai('gpt-4o-mini'),
    prompt: 'Please come up with 10 definitions for AI agents.',
    schema: z.object({
      definitions: z.array(z.string()),
    }),
  })
  console.log(result.object.definitions)
}
 
main()

We can use the describe function to help refine the generation.

import { openai } from "@ai-sdk/openai";
import { generateObject } from "ai";
import "dotenv/config";
import { z } from "zod";
 
const main = async () => {
  const result = await generateObject({
    model: openai("gpt-4o-mini"),
    prompt: "Please come up with 10 definitions for AI agents.",
    schema: z.object({
      definitions: z.array(z.string().describe("Use as much jargon as possible. It should be completely incoherent.")),
    }),
  });
  console.log(result.object.definitions);
};
 
main();