Developers

Everything you need to build with NobodyWho, the inference engine that runs LLMs fully on-device. No servers, no API keys, no hidden fees — inference happens on the user's hardware. This portal covers install commands for every platform, quickstarts, the model catalog, and a public data API for tools and agents.

Documentation GitHub OpenAPI spec

Install the SDK

NobodyWho ships as a native SDK for each platform. Pick yours and add it to your project.

Python

pip install nobodywho

Docs

React Native

TypeScript / JavaScript

npm install react-native-nobodywho

Docs Starter project

Expo

TypeScript / JavaScript

npx expo install react-native-nobodywho

Docs Starter project

Flutter

Dart

flutter pub add nobodywho

Docs Starter project

Swift

Add via Swift Package Manager.

Docs Starter project

Kotlin / Android

Kotlin

implementation("ai.nobodywho:nobodywho-android:2.2.0")

Docs

Godot

GDScript

Add via Godot Asset Library.

Docs

Quickstart

Load a GGUF model and start chatting — here in Python. See the docs for every language.

from nobodywho import Chat

chat = Chat("./model.gguf")
response: str = chat.ask("What is the capital of denmark?").completed()

# The capital of Denmark is Copenhagen.

Need a model? Browse our recommended selection or query it programmatically below. These are just recommendations — NobodyWho runs any compatible model: text generation in GGUF, and speech-to-text, text-to-speech and voice activity detection in ONNX, loaded from Hugging Face or any URL.

Data API

A public, read-only JSON API describing the model catalog, supported platforms and the blog. No authentication, CORS enabled, safe for tools and agents. Described by an OpenAPI 3.1 spec.

Endpoint	Description
GET /api/models.json	Recommended on-device model selection (NobodyWho runs any GGUF/ONNX model)
GET /api/platforms.json	Supported SDK platforms & install commands
GET /api/posts.json	Blog posts and tutorials
GET /api/index.json	API discovery document

For AI agents

NobodyWho is built to be machine-readable. Agents can discover and integrate with this site through:

/llms.txt — a concise, LLM-friendly map of the site.
/openapi.json — OpenAPI 3.1 description of the data API.
/api/index.json — the data API discovery document.
WebMCP — browser-resident agents can call in-page tools (search_models, get_install_command, list_platforms, list_blog_posts) registered via document.modelContext.