Swift Llama

An unofficial Swift client library for interacting with Meta's Llama API.

Requirements

• Swift 5.9+
• iOS 16+
• macOS 13+
• watchOS 9+
• tvOS 16+

Installation

Add the following to your Package.swift file:

Package(
    dependencies: [
        .package(url: "https://github.com/nathanborror/swift-llama", branch: "main"),
    ],
    targets: [
        .target(
            name: "YourApp",
            dependencies: [
                .product(name: "Llama", package: "swift-llama"),
            ]
        ),
    ]
)

Usage

Chat Completion

import Llama

let client = Client(apiKey: LLAMA_API_KEY)

let request = ChatRequest(
    model: "Llama-4-Scout-17B-16E-Instruct-FP8",
    messages: [
        .init(role: .system, content: [.init(text: "You are a helpful llama.")]),
        .init(role: .user, content: [.init(text: "Hello, Llama!")])
    ]
)

do {
    let response = try await client.chatCompletions(request)
    print(response.completion_message.content.text)
} catch {
    print(error)
}

List Models

import Llama

let client = Client(apiKey: LLAMA_API_KEY)

do {
    let response = try await client.models()
    print(response.data.map { $0.id }.joined(separator: "\n"))
} catch {
    print(error)
}

Command Line Interface

$ make
$ ./llama
OVERVIEW: A utility for interacting with Meta's Llama API.

USAGE: llama <subcommand>

OPTIONS:
  --version               Show the version.
  -h, --help              Show help information.

SUBCOMMANDS:
  models                  Returns available models.
  chat                    Completes a chat request.
  chat-stream             Completes a chat streaming request.

  See 'cli help <subcommand>' for detailed help.