Install

🛠 Install ATOM

ATOM is your private local AI terminal — no cloud, no surveillance, no nonsense.

✅ Requirements

  • 🐍 Python 3.10+
  • 🧠 Ollama (for local LLMs like Gemma or Mistral)
  • 🖥 Node.js + npm (for frontend UI)

🚀 Quick Start (One-Liner)

git clone https://github.com/cipherswitch/atom && cd atom && ./run.sh

🔍 Manual Setup

If you prefer setting things up step-by-step:

1. Clone the Repo

git clone https://github.com/cipherswitch/atom
cd atom

2. Backend Setup (FastAPI)

cd backend
pip install -r requirements.txt

Then run it:

uvicorn main:app --reload

3. Frontend Setup (Vite + React)

cd ../frontend
npm install
npm run dev

4. Optional: Ollama + Models

Install Ollama (opens in a new tab) and pull a model:

ollama run gemma:7b

Supported models:

  • 🧠 Gemma
  • 🧠 Mistral
  • 🧠 Mixtral
  • 🧠 Phi-2 / Phi-3
  • 🧠 LLaMA3 (experimental)

🧪 Test It Out

Once both servers are running:

  • Visit http://localhost:5173
  • Drop in a file
  • Ask ATOM something like:
    "Summarize this PDF."
    "Who is mentioned in the doc?"

🧠 Features You Just Enabled

  • File chunking + memory injection
  • Local LLM chat
  • Modular tool support (e.g., TTS, search)
  • Persona tone + streaming (optional)
  • Real-time logs + model switching
IndexMemory