Skip to content

Odibi Documentation

Declarative data engineering. YAML in, pipelines out.

⚡ Quick Start

pip install odibi

# Create a project from template
odibi init my_project --template star-schema
cd my_project

# Run it
odibi run odibi.yaml

# View the audit report
odibi story last

That's it. You now have a working star schema pipeline.

🎯 Start Here

Goal Go to
Get running in 10 minutes Golden Path
Copy THE working config THE_REFERENCE.md ⭐
Solve a specific problem Playbook

What is Odibi?

Odibi is a framework for building data pipelines. You describe what you want in YAML; Odibi handles how.

  • Declarative: YAML over imperative Python
  • Auditable: Every run generates a "Data Story" (HTML report)
  • Multi-engine: Pandas (local), Polars (high-performance), Spark (production) with zero config changes

CLI Cheat Sheet

# Create & Run
odibi init my_project          # Scaffold from template
odibi run odibi.yaml           # Execute pipeline
odibi validate odibi.yaml      # Check config without running

# Debug
odibi story last               # View most recent story
odibi story last --node X      # Inspect a specific node
odibi doctor                   # Check environment health
odibi graph odibi.yaml         # Visualize dependencies

The Canonical Example

Every new user should run this first:

cd docs/examples/canonical/runnable
odibi run 04_fact_table.yaml

This builds a complete star schema with dimensions, facts, FK lookups, and orphan handling. See the full breakdown →

Documentation Map

New to Odibi?

  1. Golden Path — Zero to running in 10 minutes
  2. THE_REFERENCE.md — The one example to copy

Building Pipelines?

Going to Production?

Last updated: March 2026