Skip to content
pubscout

nobodywho

6.0.0, released 3 weeks ago by nobodywho.ooo

Run local LLMs, vision, embeddings, RAG, tool calling, STT, and TTS on mobile and desktop.

WASM-ready
$ flutter pub add nobodywho
Source repository

What it does

Allows developers to run large language models, vision models, speech-to-text, text-to-speech, and voice activity detection completely offline and on-device using GGUF and ONNX models with Metal/Vulkan acceleration.

AI-written summary of version 6.0.0, based on the package's README and metadata. Report an inaccuracy

Use it when

  • You need offline and on-device AI inference without relying on cloud APIs.
  • Your app needs local RAG, semantic search, or tool calling capabilities.
  • You want to process speech-to-text or text-to-speech locally within the app.

Skip it when

  • You target low-end devices with limited RAM that cannot handle model sizes.
  • You want to keep your application binary size extremely small.

Strengths

  • Runs completely offline with no API keys required.
  • Supports text, vision, audio, embeddings, RAG, and tool calling in one library.
  • Hardware acceleration via Metal and Vulkan.

Weaknesses

  • Large model sizes significantly increase app download or memory footprints.
  • Performance depends heavily on the user's device specs and available RAM.

Setup

Kind
Plugin (native code)
Effort
Low
Code generation
Yes (build_runner)
Direct dependencies
4

Initialize the engine via nobodywho.NobodyWho.init() in main.dart and provide local or downloaded GGUF/ONNX models.

import 'package:nobodywho/nobodywho.dart' as nobodywho;

void main() async {
  WidgetsFlutterBinding.ensureInitialized();
  await nobodywho.NobodyWho.init();
  final chat = await nobodywho.Chat.fromPath(modelPath: './model.gguf');
  final msg = await chat.ask('Is water wet?').completed();
  print(msg);
}

Alternatives

Packages that do the same job and could replace nobodywho.

  • llamadart

    If you want a lightweight Dart-focused solution for running llama.cpp models.

  • tflite_flutter

    If you prefer deploying TensorFlow Lite models for custom ML tasks.

Works with

Packages commonly used together with nobodywho: ones it builds on, ones built on it, and its own family.