Sentiment Analysis API

Analyze sentiment in multiple languages using KrosAI’s models

The KrosAI Sentiment Analysis API analyzes text to determine sentiment and emotional tone, with robust support for African languages. This is particularly useful for understanding customer feedback, social media monitoring, and market research in low-resource languages.

Overview

The sentiment analysis API examines text content and returns a sentiment classification (positive, negative, neutral, or mixed) along with confidence scores and optional detailed analysis.

This endpoint is OpenAI compatible

Endpoint

POST https://api.krosai.com/v1/sentiment

Request Parameters

Parameter	Type	Required	Description
text	string	Yes	The text to analyze (1-5000 characters)
language	string	No	Optional language specification (e.g., "yoruba", "english", "swahili")
model	string	No	Model to use (default: "KrosMlingual2.0.1")
detailed	boolean	No	Whether to include detailed sentiment scores and emotion analysis (default: false)

Example Request

{
  "text": "I am very happy with the service provided. Thank you!",
  "language": "english",
  "detailed": true
}

Response Format

{
  "id": "sent_1234567890",
  "object": "sentiment_analysis",
  "created": 1682456789,
  "model": "KrosMlingual2.0.1",
  "results": {
    "sentiment": "positive",
    "confidence": 0.92,
    "details": {
      "positive_score": 0.92,
      "neutral_score": 0.06,
      "negative_score": 0.02,
      "emotions": {
        "joy": 0.85,
        "satisfaction": 0.78,
        "contentment": 0.65
      }
    }
  },
  "language": {
    "detected": "english",
    "confidence": 0.99
  },
  "usage": {
    "prompt_tokens": 25,
    "completion_tokens": 45,
    "total_tokens": 70
  }
}

Batch Processing

For analyzing multiple texts in a single request, use the batch endpoint:

POST https://api.krosai.com/v1/sentiment/batch

Batch Request Example

{
  "texts": [
    {
      "text": "The product works perfectly!",
      "language": "english"
    },
    {
      "text": "أنا لا أحب هذه الخدمة على الإطلاق.",
      "language": "arabic"
    }
  ],
  "detailed": true
}

This returns an array of sentiment analysis results for each text.

Supported Languages

The API specializes in African languages, including:

• Yoruba

• English

• Hausa

• Igbo

• Swahili

• Pidgin

• Arabic

• French

• Spanish

• Portuguese

• Amharic

• Zulu

• Xhosa

• Somali

• Oromo

Best Practices

Sentiment Analysis

• Provide longer text samples for more accurate sentiment analysis

• When language is known, specify it for better accuracy

• Use detailed analysis for deeper insights into sentiment

Error Handling

All APIs return standard HTTP status codes:

• 200: Success

• 400: Bad request (check parameters)

• 401: Unauthorized (check API key)

• 429: Rate limit exceeded

• 500: Server error

Error responses include a detail field with more information about the error.