Starkiller - BI Generativo com IA
jbouder/starkiller
Gere dashboards e visualizações com IA a partir de perguntas naturais
Ver no GitHubFerramenta experimental de Business Intelligence que usa LLMs para gerar visualizações de dados em tempo real a partir de consultas em linguagem natural. Em vez de dashboards pré-construídos, você faz perguntas sobre seus dados e recebe gráficos e insights visuais gerados dinamicamente pela IA.
Instalação
npx convez-ai jbouder/starkiller1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Análise exploratória de dados sem pré-definir dashboards
- Geração dinâmica de gráficos e relatórios visuais sob demanda
- Consultas conversacionais sobre métricas e KPIs de negócio
- Prototipagem rápida de visualizações de dados
Veredito

Projeto ambicioso que reimagina BI tradicional com IA gerativa, integrando Claude/Gemini com FastAPI e React. Experimento promissor para análise conversacional, mas ainda em fase conceitual — requer setup local ou Docker e chaves de API para operacional.
README
A tradução para o português deste README ainda está na fila.
Starkiller
An experimental Generative BI (Business Intelligence) tool that leverages Large Language Models to generate data visualizations on the fly.
Overview
Starkiller reimagines traditional BI dashboarding by using AI to dynamically create data visualizations based on natural language queries. Instead of pre-built dashboards, users can ask questions about their data and receive intelligent, contextual visualizations generated in real-time.
Features
- AI-Powered Visualization: Generate charts, graphs, and dashboards using natural language
- Dynamic Analysis: Ask questions and get instant visual insights
- Flexible Data Sources: Connect to various data sources and databases
- Interactive Dashboards: Explore and refine visualizations interactively
- Smart Recommendations: AI suggests relevant visualizations based on your data
Tech Stack
| Frontend | Backend |
|---|---|
| React 19 + TypeScript | Python FastAPI |
| Vite | SQLAlchemy (async) |
| Tailwind CSS | Anthropic Claude & Gemini |
| shadcn/ui | Pandas/NumPy |
| Recharts | Alembic + PostgreSQL |
Quick Start (Docker Compose)
The easiest way to run Starkiller is with Docker Compose, which starts the UI, API, and PostgreSQL database together.
Prerequisites
- Docker and Docker Compose
- An Anthropic API key or Google Gemini API key
1. Clone and Configure
git clone https://github.com/yourusername/starkiller.git
cd starkiller
# Configure the API environment
cp api/.env.example api/.env
Edit api/.env and configure your preferred LLM provider:
# Supported providers: anthropic, gemini
LLM_PROVIDER=anthropic
# If using Anthropic
ANTHROPIC_API_KEY=sk-ant-your-key-here
# If using Gemini
GEMINI_API_KEY=your-gemini-key-here
2. Start the Stack
docker compose up --build
3. Initialize the Database
In a new terminal, run migrations and seed sample data:
# Run database migrations
docker compose exec api alembic upgrade head
# Seed external sample databases
docker compose exec api python scripts/seed_connection_data.py
# Seed application data sources
docker compose exec api python scripts/seed_data_sources.py
# Seed sample dashboards
docker compose exec api python scripts/seed_dashboards.py
4. Access the Application
| Service | URL |
|---|---|
| UI | http://localhost:3000 |
| API | http://localhost:8000 |
| API Docs (Swagger) | http://localhost:8000/docs |
| API Docs (ReDoc) | http://localhost:8000/redoc |
Local Development (Without Docker)
For development without Docker, you'll need to run the frontend and backend separately.
Prerequisites
- Node.js 18+ and npm 9+
- Python 3.11+
- PostgreSQL 15+
- An Anthropic API key or Google Gemini API key
Backend Setup
cd api
# Create and activate virtual environment
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
# Install dependencies
pip install -e ".[dev]"
# Configure environment
cp .env.example .env
# Edit .env and configure your LLM provider
# Example for Gemini:
# LLM_PROVIDER=gemini
# GEMINI_API_KEY=your-key-here
# Run migrations
alembic upgrade head
# Start the API server
uvicorn main:app --reload
Frontend Setup
cd ui
# Install dependencies
npm install
# Start the development server
npm run dev
Seed Sample Data (Local)
With PostgreSQL running locally:
cd api
source .venv/bin/activate
# Set admin connection for creating external databases
export ADMIN_DB_URL=postgresql://postgres:postgres@localhost:5432/postgres
# Seed in order
python scripts/seed_connection_data.py # External sample databases
python scripts/seed_data_sources.py # Application data sources
python scripts/seed_dashboards.py # Sample dashboards
Local URLs
| Service | URL |
|---|---|
| UI | http://localhost:5173 |
| API | http://localhost:8000 |
Common Commands
Running Tests
cd api
pytest # Run all tests
pytest --cov=. --cov-report=html # With coverage report
pytest tests/unit/ # Unit tests only
pytest -k "test_name" # Run specific test
Building for Production
cd ui
npm run build # Build frontend
npm run preview # Preview production build
Adding UI Components
cd ui
npx shadcn@latest add <component-name>
Linting
cd ui
npm run lint
Environment Variables
API (api/.env)
| Variable | Description | Default |
|---|---|---|
LLM_PROVIDER | LLM provider (anthropic, gemini) | anthropic |
ANTHROPIC_API_KEY | Anthropic API key | (required if using anthropic) |
ANTHROPIC_MODEL | Claude model to use | claude-sonnet-4-5-20250929 |
GEMINI_API_KEY | Google Gemini API key | (required if using gemini) |
GEMINI_MODEL | Gemini model to use | gemini-2.0-flash |
DATABASE_URL | Database connection string | postgresql+asyncpg://postgres:postgres@localhost:5432/starkiller |
ENVIRONMENT | development, staging, production | development |
DEBUG | Enable debug mode | true |
HOST | API host | 0.0.0.0 |
PORT | API port | 8000 |
CORS_ORIGINS | Allowed CORS origins | ["http://localhost:5173"] |
License
This project is licensed under the GNU General Public License v3.0 - see the LICENSE.md file for details.