Run powerful AI models directly on-device — in seconds, with just a few lines of code.
No cloud. No API keys. No latency. Fully offline.

1nm is a collection of lightweight SDKs that make it easy to integrate local LLMs into mobile apps.

Instead of dealing with native code, model loading, and platform differences —
you get a simple, clean API that just works.

iOS support will use the most optimal runtime for the platform (not necessarily llama.cpp).

final ai = OneNm(model: OneNmModel.qwen25);
await ai.initialize();

final reply = await ai.chat('Hello!');

That’s it. You now have on-device AI running inside your app.

▶️ Watch the demo video here

• ⚡ Fast — runs directly on-device
• 🔒 Private — no data leaves the user’s phone
• 💸 Zero API cost — no tokens, no billing
• 🧩 Simple integration — minimal setup, clean APIs
• 📱 Built for mobile — not a desktop tool forced onto phones

• Model download & storage
• Native runtime integration
• Memory & performance handling
• Multi-turn chat

So you can focus on building features — not infrastructure.

onenm_local_llm/
├── flutter-llama.cpp/    # Flutter plugin (pub.dev: onenm_local_llm)
│   ├── lib/              # Dart API
│   ├── android/          # Native layer (Kotlin + C++)
│   ├── example/          # Demo apps
│   └── test/             # Tests
├── .github/
├── CONTRIBUTING.md
├── CODEOWNERS
└── LICENSE

Start with Flutter:

👉 flutter-llama.cpp/README.md

• Kotlin SDK
• iOS support
• More model options
• Performance improvements
• Additional language bindings

We welcome contributions to improve onenm_local_llm! At this stage, contributions are limited to adding new models to the model registry, testing them. and other minor improvements.

See CONTRIBUTING.md

MIT License — see LICENSE

llama.cpp is also licensed under MIT: https://github.com/ggml-org/llama.cpp/blob/master/LICENSE