@ondeinference / onde-swift
Code Commits Issues Pulls Cloud Agent CI/CD
main
feature/example main
md 1.85 KB
View source Raw

Onde example app

A one-screen SwiftUI chat app. Loads an LLM on your iPhone and streams replies, all on-device.

Setup

Install XcodeGen if you don't have it:

brew install xcodegen

From the onde-swift repository root, build the local Onde XCFramework for iOS:

make ios

This regenerates Sources/Onde/onde.swift from your local Rust build in ../onde and writes OndeFramework.xcframework into the package root so SwiftPM can link it.

Then generate the Xcode project and open it:

cd Examples/OndeExample
xcodegen generate
open OndeExample.xcodeproj

Set your development team under Signing & Capabilities, pick your iPhone or simulator, hit Run.

What it does

On first launch, it downloads and loads the default model for your device (Qwen 2.5 Coder 1.5B on iPhone, 3B on Mac). After that you can chat with it.

Current example app features:

  • token-by-token streaming with a stop button
  • local conversation persistence between launches
  • model settings sheet for system prompt + sampling preset
  • toolbar actions to reload the model or clear the conversation
  • prompt suggestions when the conversation is empty

Files

File What's in it
App.swift Entry point. Calls setupInferenceEnvironment() at launch.
ContentView.swift The chat UI: header, empty state, message list, input bar, reload/clear actions.
ChatViewModel.swift Bridges UniFFI streaming into Swift concurrency, manages model lifecycle, restores history, persists conversation.
ExampleSupport.swift Prompt suggestions, sampling presets, and the settings sheet.
InferenceEnvironment.swift Sets HF_HOME inside the app sandbox so HuggingFace downloads land somewhere iOS allows.
project.yml XcodeGen spec. The .xcodeproj is generated from this, not checked in.