Examples
Here are some real-world application examples of AI for your codebase.
These are just some useful examples. Please feel free to add more via Pull Requests.
Full Chat (Demo)
<script setup lang="ts">
const { messages, send } = useGPT().createChat()
const input = ref<string>('')
</script>
<template>
<pre>{{ messages }}</pre>
<input
type="text"
v-model="input"
@keyup.enter="send(input, () => (input = ''))" />
</template>
Single message
const { answer } = useGPT('Give me 5 alternative search terms for sneakers').send()
Check spelling
const res = await useGPT('Check spelling of "kaola bear" and correct it if neccessary')
.toStructured(
z.object({
isCorrect: z.boolean(),
correctedText: z.string()
})
)
// Output
{
isCorrect: false,
correctedText: 'koala bear'
}
Alternative search
const response = useGPT('Provide 5 alternative search terms')
.addContextData('sneakers')
.toStructured(
z.object({
suggestions: z.array(z.string()),
}),
)
// Output
{
suggestions: [
'trainers',
'athletic shoes',
'running shoes',
'casual sneakers',
'sport shoes'
]
}
Note: This example uses
.addContextData for the base term. The term could of course also be added in the prompt itself.Read article aloud
const audioURLBase64 = await useGPT(articleRef.value.textContent)
.toSpeech()
Get Weather
const res = await useGPT('Current weather in San Francisco')
.toStructured(
z.object({
celsius: z.number(),
fahrenheit: z.number(),
weatherLabel: z.object({
en: z.string(),
de: z.string(),
})
})
)
// Output
{
celsius: 15,
fahrenheit: 59,
weatherLabel: {
en: "Partly Cloudy"
de: "Teilweise bewölkt",
}
}