Inference API

Inference API

This guide covers how to make requests to the Onairos Inference API and understand the responses.

Making API Requests

When autoFetch is Disabled

If you set autoFetch = false, you'll receive the API URL and token via window.postMessage:

useEffect(() => {
  const handleMessage = (event) => {
    if (event.data?.source === 'content-script' && 
        event.data?.type === 'API_URL_RESPONSE') {
      const { apiUrl, accessToken } = event.data;
      fetchOnairosData(apiUrl, accessToken);
    }
  };
  
  window.addEventListener('message', handleMessage);
  return () => window.removeEventListener('message', handleMessage);
}, []);

Request Format

Field	Type	Required	Description
text	String	Yes	Text input for inference
category	String	Yes	Category of the content
img_url	String	No	URL of associated image

const response = await fetch(apiUrl, {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'Authorization': `Bearer ${accessToken}`
  },
  body: JSON.stringify({
    Input: {
      item1: { text: "Product description", category: "fashion" },
      item2: { text: "Another item", category: "lifestyle", img_url: "https://..." }
    }
  })
});

const data = await response.json();

---

Response Format

Sentiment Scores

Sentiment predictions return scores between 0-1 indicating how much the user would like each input item:

{
  "output": [
    [[0.9998]],
    [[0.0013]]
  ]
}

Score	Meaning
0.9+	User will highly engage/like this content
0.5-0.9	Moderate positive response
<0.5	User likely won't engage

Personality Traits

If the user authorized trait sharing:

{
  "Traits": {
    "positive_traits": {
      "creativity": 9.7,
      "empathy": 9.6,
      "curiosity": 9.1
    },
    "traits_to_improve": {
      "patience": 1.7,
      "organization": 3.6
    }
  }
}

Traits are scored 0-10. Use these to personalize experiences, recommendations, and interactions.