runanywhereai/runanywhere-sdks

★ 10,350C++Audience · developerComplexity · 4/5Setup · hard

Mindmap

mindmap
  root((runanywhere-sdks))
    Capabilities
      Text generation
      Speech to text
      Text to speech
      Image generation
    Platforms
      iOS Swift
      Android Kotlin
      Browser JS
      React Native
      Flutter
    How it works
      On-device only
      No internet needed
      Load model file
    Audience
      Mobile developers
      App builders
      Privacy-first devs

mindmap root((runanywhere-sdks)) Capabilities Text generation Speech to text Text to speech Image generation Platforms iOS Swift Android Kotlin Browser JS React Native Flutter How it works On-device only No internet needed Load model file Audience Mobile developers App builders Privacy-first devs

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Build a private voice assistant app for iOS that listens, transcribes, and responds entirely on the device without sending audio to a server.

USE CASE 2

Add offline text generation to an Android app so users can chat with an AI assistant without an internet connection.

USE CASE 3

Embed image generation into a React Native app so users can create pictures from text prompts inside your app.

USE CASE 4

Create a cross-platform Flutter app with on-device speech-to-text transcription using the Whisper model.

Tech stack

C++SwiftKotlinJavaScriptReact NativeFlutter

Getting it running

Difficulty · hard Time to first run · 1h+

Model files like SmolLM or Whisper are downloaded separately and can be several hundred megabytes each, browser, React Native, and Flutter SDKs are still in beta.

In plain English

RunAnywhere is a set of software development kits (SDKs) that let app developers add AI features to their applications in a way that runs entirely on the user's device. Nothing is sent to a cloud server. The AI processing happens locally, which means it works without an internet connection and user data stays private. The four main capabilities are text generation (chat with an AI assistant), speech-to-text (transcribing spoken audio), text-to-speech (having the AI speak a response aloud), and image generation (creating pictures from text descriptions). These can be combined: for example, a voice assistant that listens to the user, processes the speech, generates a reply with a language model, and speaks the response back, all offline. SDKs are available for iOS and macOS using Swift, Android using Kotlin, browsers using JavaScript, React Native for cross-platform mobile development, and Flutter. The Swift and Kotlin versions are listed as stable. The browser, React Native, and Flutter versions are in beta. Each SDK follows the same basic three-step pattern: initialize the library, load a model file, then call the generation function. Model files such as SmolLM, Qwen, or Whisper are downloaded separately. The repository also includes complete sample applications for each platform so developers can see a working implementation before writing their own. There is a companion consumer app on the iOS App Store and Google Play that demonstrates the capabilities. This is aimed at software developers building apps, not end users looking for a standalone AI tool. Integrating it requires writing code in the relevant platform's language and managing model files, which can be several hundred megabytes each. The README is longer than what was shown.

Copy-paste prompts

Prompt 1

I want to add an offline AI chat feature to my iOS app using the RunAnywhere Swift SDK. Show me the three-step initialization, model loading, and generation code using the SmolLM model.

Prompt 2

Help me integrate the RunAnywhere Kotlin SDK into my Android app to transcribe voice input using Whisper without sending audio to a cloud service.

Prompt 3

I'm building a React Native app and want to use the RunAnywhere beta SDK for on-device text generation. Walk me through setup and show a working example component.

Prompt 4

What are the tradeoffs between using SmolLM vs Qwen as the language model in the RunAnywhere SDK, and how do I decide which to use for a mobile app?

Open on GitHub → Explain another repo

← runanywhereai on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.