Quick Start

import RunAnywhere
import RunAnywhereLlamaCPP

LlamaCPP.register()
try RunAnywhere.initialize(apiKey: "", baseUrl: "", environment: .development)

let result = try await RunAnywhere.llm.generate(
    prompt: "Explain quantum computing in one paragraph",
    options: LlmOptions(model: "qwen3-0.6b", maxOutputTokens: 200)
)

print(result.text)
print("\(result.tokensPerSecond) tok/s")

for try await event in try await RunAnywhere.llm.generateStream(prompt: "Write a haiku") {
    switch event {
    case .token(let token):
        print(token.text, terminator: "")
    case .completed(let result):
        print("\n\(result.tokensPerSecond) tok/s")
    default:
        break
    }
}

let messages = [
    ChatMessage(role: .system, content: "You are a concise assistant."),
    ChatMessage(role: .user, content: "What is on-device inference?"),
    ChatMessage(role: .assistant, content: "Running the model on your own hardware."),
    ChatMessage(role: .user, content: "Why does that matter?"),
]

let reply = try await RunAnywhere.llm.generate(messages: messages)

import RunAnywhereONNX
ONNX.register()

// Speech to text
let transcription = try await RunAnywhere.stt.transcribe(AudioInput.file(path))
print(transcription.text)

// Text to speech
let handle = try await RunAnywhere.tts.speak("Ready when you are")
await handle.waitForPlayout()

import RunAnywhere
import RunAnywhereLlamaCPP
import SwiftUI

@MainActor
@Observable
final class ChatModel {
    var reply = ""
    var isGenerating = false

    func boot() throws {
        LlamaCPP.register()
        try RunAnywhere.initialize(apiKey: "", baseUrl: "", environment: .development)
    }

    func send(_ prompt: String) async {
        isGenerating = true
        reply = ""
        defer { isGenerating = false }

        do {
            let stream = try await RunAnywhere.llm.generateStream(prompt: prompt)
            for try await event in stream {
                if case .token(let token) = event { reply += token.text }
            }
        } catch {
            reply = "Failed: \(error.localizedDescription)"
        }
    }
}

struct ChatView: View {
    @State private var model = ChatModel()
    @State private var prompt = ""

    var body: some View {
        VStack {
            ScrollView { Text(model.reply).frame(maxWidth: .infinity, alignment: .leading) }
            HStack {
                TextField("Ask something", text: $prompt)
                Button("Send") { Task { await model.send(prompt) } }
                    .disabled(model.isGenerating || prompt.isEmpty)
            }
        }
        .padding()
        .task { try? model.boot() }
    }
}