Kal-AI

A fully local multi-agent AI system powered by Ollama — no API costs, no cloud. Built with Python (FastAPI) and a pixel art office UI.

Prerequisites

1. Python 3.11+ — python3 --version
2. Ollama — https://ollama.ai
3. Pull a model:

ollama pull llama3.1

Setup

git clone https://github.com/kale87/kal-ai
cd kal-ai
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
# Edit .env — add your GITHUB_TOKEN
python main.py

Open http://localhost:3000

Agents

Agent	Role
🎯 Manager	Understands your request, delegates to specialists, synthesizes results
💻 Coder	Code writing, review, debugging, GitHub operations
🔍 Researcher	Research, summarizing, context gathering
✍️ Writer	Content, documentation, commit messages

Switching models

Edit OLLAMA_MODEL in .env and restart. Any model pulled with ollama pull works.