NobodyWho Flutter Starter App
A Flutter starter app demonstrating NobodyWho, a library for running LLMs locally and efficiently on any device, covering chat, tool calling, multimodal vision and hearing, embeddings and RAG, speech-to-text and text-to-speech.
This starter app shows how to use NobodyWho, a library for running large language models locally on-device, from a Flutter project. Demonstrated features include streaming chat responses from a local model, tool calling that gives the model access to custom functions such as a weather lookup or calculator, image and audio ingestion with a multimodal model, semantic search and retrieval-augmented generation using an embedding model with a cross-encoder reranker, speech-to-text transcription, and text-to-speech generation.
Models can be downloaded ahead of time into the app's assets for development, or fetched on demand in production using a background downloader or NobodyWho's own hf:// download helper, which caches Hugging Face models locally and reuses them on later loads.
The app has been tested on iOS, Android and macOS, and is expected to also work on Linux and Windows; Flutter Web is not supported.