The Reactive Notebook That Ships to Production
FlowyML Notebook is a reactive, DAG-powered notebook environment that replaces Jupyter for production ML workflows. Write pure Python cells, get automatic dependency tracking, and ship directly to pipelines, dashboards, and apps — without changing a single line of code.
Reactive notebook editor with code cells, variable explorer, and full toolbar
| Feature | Jupyter | Deepnote | Marimo | FlowyML Notebook |
|---|---|---|---|---|
| Reactive DAG Execution | ❌ | ❌ | ✅ | ✅ |
Pure .py File Storage |
❌ | ❌ | ✅ | ✅ |
| Git-Native Collaboration | ❌ | ❌ | ✅ GitHub | |
| Pipeline Integration | ❌ | ❌ | ❌ | ✅ FlowyML |
| Reusable Recipes | ❌ | ❌ | ❌ | ✅ |
| One-Click Deploy | ❌ | ❌ | ✅ | |
| SQL First-Class | ❌ | ✅ | ✅ | ✅ |
| AI Assistant | ❌ | ✅ | ❌ | ✅ |
| Rich Data Explorer | ❌ | ✅ | ✅ | ✅ |
| App Mode | ❌ | ❌ | ✅ | ✅ |
| Self-Hosted | ✅ | ❌ | ✅ | ✅ |
| SmartPrep Advisor | ❌ | ❌ | ❌ | ✅ |
| Algorithm Matchmaker | ❌ | ❌ | ❌ | ✅ |
| Interactive Dashboards | ❌ | ❌ | ❌ | ✅ |
| Analysis Patterns | ❌ | ❌ | ❌ | ✅ |
| Keras Ecosystem | ❌ | ❌ | ❌ | ✅ UnicoLab |
# Install the core package
pip install flowyml-notebook
# Or install with all ML & AI extensions
pip install "flowyml-notebook[all]"
# Or install with Keras ecosystem (KDP + KerasFactory + MLPotion)
pip install "flowyml-notebook[keras]"fml-notebook dev # 🔥 Hot-reload development mode
fml-notebook start # 🚀 Production build💡 Short alias: You can also use
fmlninstead offml-notebook— it's the same CLI:fmln dev # same as fml-notebook dev fmln start # same as fml-notebook startYou can also import the package with
import fmlnas a shorthand forimport flowyml_notebook.
The browser opens automatically. You're ready to build.
Every DataFrame gets automatic profiling — statistics, distributions, correlations, quality checks, and ML-ready insights. No extra code needed.
Automatic DataFrame profiling with column statistics, type detection, and memory impact
Interactive charts for every column — histograms, bar charts, and distribution analysis
Pearson correlation matrix with color-coded heatmap for quick feature analysis
Automated ML insights: outlier detection, scaling recommendations, and target variable suggestions
Cells are nodes in a dependency graph. Change a variable, and only dependent cells re-execute — automatically. Visualize the full pipeline with the built-in DAG view.
Visual dependency graph showing data flow: imports → data_generation → analysis → exploration → summary
Native integration with the UnicoLab ML ecosystem — KDP, KerasFactory, and MLPotion. All packages are optional and auto-detected.
| Package | What It Does | Integration Point |
|---|---|---|
| KDP | Keras preprocessing layers with distribution-aware encoding | SmartPrep Advisor |
| KerasFactory | 38+ reusable Keras layers and production-ready model architectures | Algorithm Matchmaker |
| MLPotion | Managed training pipelines with type-safe configuration | Algorithm Matchmaker |
When all three are installed, the Algorithm Matchmaker surfaces a flagship end-to-end pipeline: KDP → KerasFactory → MLPotion — preprocessing, model building, and training in a single deployable Keras model.
# Install the full ecosystem
pip install "flowyml-notebook[keras]"4 new builtin recipes are also included: KDP Smart Preprocessing, KerasFactory Quick Model, MLPotion Training Pipeline, and the UnicoLab End-to-End Pipeline.
Stop rewriting boilerplate. 43 built-in recipes across Core, Assets, Parallel, Observability, Evals, Data, ML, Visualization, and Ecosystem categories. Drag into your notebook or click to insert.
Searchable recipe library with FlowyML Step, Pipeline, Conditional Branching, and more
Collaborate directly in the notebook with inline comments and a review panel. Add notebook-level or cell-level annotations for team discussions.
Comments panel with threaded discussions, resolve/reply actions, and scatter plot output
Generate beautiful HTML or PDF reports from your notebook. Optionally include source code cells alongside outputs. Preview in browser, then download.
Report generation with HTML/PDF format selection, code inclusion toggle, and instant preview
Turn any notebook into an interactive web application with one click. Choose layout (Linear, Grid, Tabs, Sidebar, Dashboard), theme, and cell visibility.
Publish dialog with layout options, dark/light/auto theme, source code toggle, and per-cell visibility
Ship notebooks directly to production. Promote to pipeline, deploy as API/Docker/Batch, track kernel assets (DataFrames, models), and connect to FlowyML infrastructure.
Pipeline promotion with quick actions and @step decorators |
Deploy as API, Docker Container, or Batch Pipeline with infrastructure stacks |
Kernel assets: tracked DataFrames with size, shape, and type metadata
Full GitHub integration as the collaboration backend. Link a repository, branch, commit, and push — all from the notebook sidebar. No proprietary cloud needed.
Connect GitHub repository for team collaboration and versioning |
Save and browse notebook snapshots with cell-level diffs |
Run standalone (Local Mode) or connect to a FlowyML server (Remote Mode) for experiment tracking, pipeline export, and deployment. Full runtime details at a glance.
Environment panel: Local/Remote connection, runtime info (Python 3.12, IPython, Reactive DAG engine)
| Command | Description |
|---|---|
fml-notebook dev |
🔥 Launch with Vite hot reload |
fml-notebook start |
🚀 Launch with production build |
fml-notebook run <file> |
|
fml-notebook export <file> |
📦 Export as pipeline/HTML/PDF/Docker |
fml-notebook app <file> |
🌐 Deploy as interactive web app |
fml-notebook list --server <URL> |
📚 List notebooks on a server |
fml-notebook convert <file> |
🔄 Convert between .ipynb ↔ .py |
fml-notebook diff <a> <b> |
📝 Compare two notebooks |
All commands also work with the short alias:
fmln dev,fmln start, etc.
FlowyML Notebook ships with 13 built-in analysis tools — a full data science toolkit embedded directly in the notebook. No extra installs, no configuration.
| Tool | What It Does |
|---|---|
| Cell Profiler | CPU time, wall time, memory delta, peak memory, and hot function detection for any cell |
| Cell Benchmark | Statistical timing (mean, median, std, min, max) with regression detection across runs |
| Tool | What It Does |
|---|---|
| Data Validator | Automatic DataFrame quality scoring — null detection, duplicate analysis, column-level diagnostics |
| Data Lineage | Track how DataFrames transform across cells — row/column changes, dtype mutations, null diffs |
| Tool | What It Does |
|---|---|
| Code Analyzer | Smart linting for data science — pandas anti-patterns, security checks, naming conventions, auto-fix |
| Cell Dependencies | AST-based dependency graph between cells — topological sort, stale detection, optimal execution order |
| Notebook Search | Full-text search with fuzzy/regex matching, find & replace, variable/function finder, duplicate detection |
| Tool | What It Does |
|---|---|
| Snippets Library | 35 built-in code snippets across 8 categories — Data Loading, EDA, Modeling, Visualization, and more |
| Execution History | Time-travel debugging — inspect past outputs, compare runs, track performance trends |
| Environment Manager | Full environment snapshot, in-notebook package install/uninstall, requirements.txt export |
| Tool | What It Does |
|---|---|
| Jupyter Import/Export | Seamless .ipynb ↔ .py conversion with full output/metadata preservation |
| Notebook Diff | Cell-level diff between notebooks with added/removed/modified detection |
| Package Installer | pip-compatible package management without leaving the notebook |
Visit FlowyML Notebook Docs for the complete guide:
- Getting Started — Install, launch, configure
- Features — Complete feature inventory
- Architecture — Reactive DAG engine internals
- Recipes — Reusable cell templates
- Collaboration — GitHub-based team workflows
- Integration — FlowyML instance connections
- Ecosystem — UnicoLab Keras ecosystem (KDP, KerasFactory, MLPotion)
- Data Exploration — Rich DataFrame profiling
- API Reference — CLI & Python API docs
git clone https://github.com/UnicoLab/flowyml-notebook.git
cd flowyml-notebook
make setup
make dev| Target | Description |
|---|---|
make setup |
🔧 Install Python package + frontend deps |
make dev |
🔥 Launch dev mode with hot reload |
make test |
🧪 Run all tests |
make lint |
🔍 Run Ruff linter |
make format |
✨ Auto-format code |
make docs |
📖 Build MkDocs documentation |
make docs-serve |
👁️ Preview docs locally |
make pre-commit |
🔒 Run pre-commit checks |
make release-dry-run |
🏷️ Dry-run semantic release |
make clean |
🧹 Remove build artifacts |
See CONTRIBUTING.md for the full contributor guide.
- 📖 Documentation
- 🐛 Bug Reports
- 💬 Discussions
- 📋 Contributing Guide
- 📜 Changelog
- 🔒 Security Policy
- 📝 Code of Conduct
Licensed under the Apache License 2.0.
Built with ❤️ by UnicoLab