⚡ Getting Started

Get up and running with FlowyML Notebook in under 60 seconds. No Docker, no cloud signup, no configuration needed.

Installation

pippoetry

# Recommended: install with all ML & AI extensions
pip install "flowyml-notebook[all]"

poetry add "flowyml-notebook[all]"

Minimal Install

Just need the core notebook? Run pip install flowyml-notebook — you can add extras later.

What's Included

Package	What You Get
Core	Reactive DAG engine, editor, recipes, reports, app publishing
`[ai]`	AI assistant (OpenAI & Google Generative AI)
`[sql]`	SQL cells with DuckDB & SQLAlchemy
`[exploration]`	Advanced DataFrame profiling with ML insights
`[all]`	Everything above — recommended

Launch

1	`fml-notebook start`

That's it. Your browser opens to http://localhost:8880 and you're ready to build.

FlowyML Notebook Editor — The FlowyML Notebook editor — reactive cells, variable explorer, and integrated toolbar

CLI Options

fml-notebook start --port 9000      # Custom port
fml-notebook start --no-browser     # Don't auto-open browser
fml-notebook dev                    # Hot-reload mode (for contributors)

Your First Reactive Notebook

1. Start with Data

Click Demo in the toolbar to load a full end-to-end example, or add cells manually:

# Cell 1 — Generate ML experiment data
import pandas as pd
import numpy as np

df = pd.DataFrame({
    'model': np.random.choice(['XGBoost', 'LightGBM', 'SVM', 'Neural Net'], 100),
    'accuracy': np.random.uniform(0.72, 0.98, 100).round(4),
    'f1_score': np.random.uniform(0.68, 0.96, 100).round(4),
    'training_time_s': np.random.exponential(30, 100).round(2),
})
df

2. Get Instant Insights

Every DataFrame automatically gets rich profiling — statistics, charts, correlations, and ML recommendations:

Automatic Data Profiling — Charts tab: histogram distributions for every column, automatically generated

ML Insights — Insights tab: outlier detection, scaling recommendations, and target variable identification

3. Experience Reactivity

# Cell 2 — This re-executes automatically when Cell 1 changes!
summary = df.groupby('model').agg(
    avg_accuracy=('accuracy', 'mean'),
    avg_f1=('f1_score', 'mean'),
    count=('model', 'count'),
).round(4).sort_values('avg_accuracy', ascending=False)
summary

Change anything in Cell 1 → Cell 2 automatically re-executes.

That's the power of the reactive DAG engine. No stale state. No "restart and run all."

4. Visualize Dependencies

Open the DAG tab to see how your cells connect:

DAG View — Visual dependency graph: imports → data_generation → analysis → exploration → summary

5. Ship to Production

When you're happy with your analysis:

Export Report — Generate HTML/PDF with one click
Publish as App — Turn it into an interactive web dashboard
Promote to Pipeline — Convert to a production FlowyML pipeline

Publish as App — Choose layout, theme, and cell visibility — then publish with one click

Connect to FlowyML (Optional)

For experiment tracking, model registry, and pipeline deployment, connect to a FlowyML instance:

Environment — Local mode works standalone — switch to Remote for full platform features

fml-notebook start --server https://flowyml.your-org.ai

Or configure from the Env panel in the sidebar.

What's Next?

Explore	Description
Features	Complete visual tour of every capability
Recipes	39 built-in code templates for ML workflows
Collaboration	GitHub-powered team workflows
Integration	Pipelines, deploy, and asset management
Data Exploration	Deep dive into DataFrame profiling

Need Help?