onLM — Private AI & LLM Chat Offline

onLM provides a secure and private environment for interacting with large language models and performing speech recognition directly on your iPhone or iPad. Key features include: • Fully on-device processing, no cloud or API keys • Choose between Apple Intelligence and open-source models (Hugging Face, MLX) • Voice transcription with Qwen3-ASR, completely offline • Record, transcribe, and summarize content locally • No subscriptions, download models once This application is built for complete privacy, ensuring your conversations and voice data never leave your device. All processing happens locally on Apple Silicon, making it an ideal choice for sensitive information or use in environments without internet access. You can select from various optimized 4-bit quantized models like Qwen, Gemma, Llama, Phi, and Mistral, each designed for efficient mobile performance. onLM empowers users with powerful linguistic capabilities wherever they go, without compromising data security. It's perfect for individuals who require robust text generation and transcription features for personal notes, secure communications, or general productivity, all while maintaining absolute control over their data. Engineers and researchers can also benefit from experimenting with different models in a sandboxed, offline setting.