OneAI logo Local AI Hub

Local AI Hub — Yours for the voyage.

Your own private AI hub,
running on your hardware.

Local AI Hub is a self-hosted interface for talking to language models — chat, documents, model management, realtime streaming, and Docker-powered services. Local-first, privacy-first, no telemetry. Just you and your models.

(if you want to wire in an external API too — that's fine, I'm not here to tell you how to live your life…)

Start installation Live demo

Why it matters

Local AI should be usable by people who are not infrastructure experts.

Privacy first

Conversations, documents, and model traffic can stay inside your own machine or network instead of being sent to a hosted AI provider by default.

Lower friction

The project turns model management, chat, document upload, and server configuration into a guided web experience instead of scattered terminal commands.

Open foundation

It uses common building blocks: Next.js, FastAPI, PostgreSQL with pgvector, Redis, Socket.IO, and Docker Compose.

What it solves

A single workspace for local models, chat, and document context.

01

Connect local inference nodes such as Ollama and keep installed models visible in the UI.

02

Chat with realtime streamed responses through Socket.IO instead of waiting on a static request.

03

Upload documents and use pgvector-backed retrieval so answers can include your own context.

04

Use optional web search when current information is more important than model memory.

Live Demo

Experience the speed of local inference.

U

Installation

Install once, then run the API and web app locally.

Windows

  1. Install Git for Windows.
  2. Install Docker Desktop and enable WSL 2 integration.
  3. Install Python 3.11 or newer.
  4. Install Node.js 20 LTS or newer.
  5. Install Ollama for Windows if you want local model hosting.
git clone <your-repo-url>
cd oneai
docker compose -f infra/compose/docker-compose.dev.yml up -d
cd backend
py -3 -m venv venv
venv\Scripts\activate
pip install -r requirements.txt
alembic upgrade head
python run.py
cd frontend
npm install
npm run dev

macOS

  1. Install Homebrew if it is not already installed.
  2. Install Docker Desktop for Mac.
  3. Install Python, Node.js, and Git.
  4. Install Ollama if you want to run local models.
brew install git python node
git clone <your-repo-url>
cd oneai
docker compose -f infra/compose/docker-compose.dev.yml up -d
cd backend
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
alembic upgrade head
python run.py
cd frontend
npm install
npm run dev

Linux

  1. Install Git, Python 3, venv, Node.js 20 LTS or newer, and Docker Engine.
  2. Start Docker and make sure your user can run Docker commands.
  3. Install Ollama if you want local inference on the same machine.
sudo apt update
sudo apt install -y git python3 python3-venv python3-pip nodejs npm
git clone <your-repo-url>
cd oneai
docker compose -f infra/compose/docker-compose.dev.yml up -d
cd backend
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
alembic upgrade head
python run.py
cd frontend
npm install
npm run dev

Run and verify

Open the web UI and connect your first local model server.

1. Services

PostgreSQL with pgvector runs on port 5433. Redis runs on 6379.

2. Backend

Start from backend/ with python run.py. API runs at http://localhost:4000.

3. Frontend

Start from frontend/ with npm run dev. Web app runs at http://localhost:3000.

4. Embeddings

Install nomic-embed-text on your active local server to use document upload and RAG.