Xcode

Xcode 16+ supports configuring a custom OpenAI-compatible endpoint for its coding assistant. By pointing it at OpenDunes, you can use any model in the catalog — including Anthropic, Google, and open-source models — for code completions and inline assistance, with usage billed in Algerian Dinars.

Configuration#

1. Open Xcode and go to Settings → AI.
2. Under Model Provider, select Custom (OpenAI-compatible).
3. Enter the following:
   • Base URL: https://opendunes.com/v1
   • API Key: your key from your dashboard
   • Model: the provider/model slug you want to use (e.g. anthropic/claude-sonnet-5)
4. Click Verify to confirm connectivity, then Save.

Choosing a Model#

For coding tasks, prefer models with strong code capabilities. Some good starting points:

Browse the full catalog at /models and filter by capability.

Swift Package — Calling OpenDunes Directly#

You can also call OpenDunes from Swift code in your app using URLSession:

import Foundation

struct Message: Codable {
    let role: String
    let content: String
}

struct ChatRequest: Codable {
    let model: String
    let messages: [Message]
}

struct ChatChoice: Codable {
    struct Delta: Codable { let content: String? }
    let message: Message
}

struct ChatResponse: Codable {
    let choices: [ChatChoice]
}

func askOpenDunes(prompt: String) async throws -> String {
    let url = URL(string: "https://opendunes.com/v1/chat/completions")!
    var request = URLRequest(url: url)
    request.httpMethod = "POST"
    request.setValue("Bearer \(ProcessInfo.processInfo.environment["OPENDUNES_API_KEY"] ?? "")", forHTTPHeaderField: "Authorization")
    request.setValue("application/json", forHTTPHeaderField: "Content-Type")

    let body = ChatRequest(
        model: "anthropic/claude-sonnet-5",
        messages: [Message(role: "user", content: prompt)]
    )
    request.httpBody = try JSONEncoder().encode(body)

    let (data, _) = try await URLSession.shared.data(for: request)
    let response = try JSONDecoder().decode(ChatResponse.self, from: data)
    return response.choices.first?.message.content ?? ""
}

Note. Never hardcode your API key in source code. Use environment variables for development and a secure secrets manager (like Xcode's .xcconfig + a server-side proxy) for production app builds.

Notes#

• Xcode's AI features require macOS 14.0+ and Xcode 16+.
• The coding assistant feature sends your current file context as part of the request — each interaction consumes tokens billed from your DA balance.
• There is no cap on the number of requests. If Xcode's assistant stops answering, check your balance and the key's spending cap on your dashboard — a spent cap or an empty wallet answers 402.