Experimenter Analytics Tools — A/B testing platform for analysts and data scientists who need full control over their experimentation workflow.
One person, full control — Own the entire workflow from setup to insights using familiar tools (Python, SQL, YAML) without requiring a separate engineering team.
Flexibility over perfection — Real experiments rarely fit templates. Expanto prioritizes adaptability, letting you modify metrics, logic, and queries as needed.
- Experiment management — Create, track, and close experiments
- Flexible calculations — Define custom metrics and pipelines.
- Statistical analysis — Built-in statistical tests and confidence intervals
- AI assistant — Multi-agent system for experiment design and analysis (it is an experiment for now).
- Web interface — Streamlit-based UI for all operations
# Clone and setup
git clone https://github.com/dmitrkozlovsk/expanto.git
cd expanto
make setup
# Configure your data warehouse
# Edit .streamlit/secrets.toml with your credentials
# Start application
make startThis will:
- Install dependencies via uv
- Copy configuration templates into .streamlit
- Create example experiment in SQLite database
- Start Streamlit UI and AI assistant
- Configuration Guide — Post-installation setup: data warehouse connections, AI assistant, secrets management
- Architecture Overview — System design, components, data flow, and technology stack
- Query Templates — Create SQL templates for your data warehouse and experiment calculations
- Metrics Configuration — Define experiment metrics in YAML: averages, ratios, proportions
- Metrics Examples — Example YAML metric definitions
- Query Examples — Example SQL templates for BigQuery and Snowflake