Installation

doesn’t matter if you’re a python nerd poking through every file or an ops guy who just wants it running — brainz installs clean, fast, and stays fully yours. no hidden deps, no cloud junk.

---

system requirements

minimum:

• python 3.10+

• node.js 18+

• postgresql 13+

• docker & docker compose (if you’re not into manual pain)

recommended:

• linux or macos (windows? you’re on your own)

• 8gb+ ram if you’re training live

• gpu (cuda) if you’re pushing big transformer weights locally

---

1. clone the repo

git clone https://github.com/brainzmonster/OS.git
cd OS

---

2. manual install (for devs who like to break stuff)

step through every layer, see what’s going on under the hood.

backend setup

cd backend
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

env setup

cp .env.example .env

example:

MODEL_NAME=tiiuae/falcon-rw-1b
DATABASE_URL=postgresql://user:pass@localhost:5432/brainz
DEBUG=true

frontend setup

cd ../web
npm install

start everything

backend:

cd ../backend
bash ../scripts/run_server.sh

frontend:

cd ../web
npm run dev

initialize db:

python ../database/db_init.py

---

3. docker (the degen way – recommended)

one command, zero headaches. docker wipes all version hell.

cp .env.example .env
docker-compose up --build

access points

• app → http://localhost

• api → http://localhost:8000

• ui → http://localhost:3000

• docs → http://localhost:8000/docs

logs stream straight to your console. tweak docker-compose.yml if you want to mount volumes or push logs somewhere fancy.

---

4. backend layout

brainz/backend/
├── api/         # rest endpoints
├── core/        # engine boot + registry
├── models/      # inference + fine-tune logic
├── agents/      # autotrain + feedback bots
├── db/          # tables + migrations
├── services/    # memory, analytics, users
├── data/        # tokenizers + embeddings
├── utils/       # metrics + crypto helpers

everything is modular. swap it, rewrite it, or delete half of it if you want — brainz won’t cry.