PerplexityApiSwift

PerplexityApiSwift is a Swift framework that provides a convenient wrapper for the Perplexity AI API. This framework simplifies the process of making chat completion requests to Perplexity's advanced language models.

Features

• Easy-to-use Swift interface for the Perplexity AI API
• Support for multiple Perplexity AI models
• Asynchronous API calls using Swift's modern concurrency features
• Built-in error handling for common API issues

Basic usage

To use PerplexityApiSwift, you need to create an instance of PerplexityAPI with your API token and then make chat completion requests. Here's a basic example:

import PerplexityApiSwift

// Initialize the API client
let api = PerplexityAPI(token: "your_api_token_here")

// Create a message
let messages = [Message(role: "user", content: "What is the capital of France?")]

// Make a chat completion request
do {
    let response = try await api.chatCompletion(messages: messages, model: .sonar)
    print(response.choices.first?.message.content ?? "No response")
} catch {
    print("Error: \(error)")
}

Important: You need to obtain an API token from Perplexity AI to use this framework. Make sure to keep your token secure and never share it publicly.

Available Models

The framework supports various Perplexity AI models through the PerplexityModel enum:

Research and Reasoning Models

• .sonarDeepResearch: Advanced research model with 128K context length
• .sonarReasoningPro: Enhanced reasoning model with 128K context length
• .sonarReasoning: Base reasoning model with 128K context length

General Purpose Models

• .sonarPro: Professional model with 200K context length
• .sonar: Standard model with 128K context length
• .r1_1776: Base model with 128K context length

Error Handling

PerplexityApiSwift defines a PerplexityError enum for common errors:

• .tokenNotSet: The API token has not been set
• .invalidResponse(statusCode:): The API returned an invalid response with the given status code
• .invalidResponseFormat: The API response could not be decoded

Upcoming Features

The following features are planned for future releases:

• Structured Outputs: Support for receiving structured, typed responses from the API
• Streaming Response: Real-time streaming of model responses for improved user experience

Documentation

For more detailed information about the Perplexity AI API, please refer to the official documentation:

Perplexity AI API Documentation