Use YAML, JSON, JSONC, TOML, and ENV directly in VS Code Jupyter notebooks without turning configuration into noisy Python boilerplate.
Config cells stay readable as config, then Convert to Variable creates and runs a hidden Python helper cell below the source cell. Your original notebook text is not rewritten, Python diagnostics do not mark config as broken Python, and the helper prints a clear variable=value confirmation.
Why
Jupyter notebooks are excellent for exploration, agent workflows, data tasks, and demos, but configuration usually ends up as fragile Python dictionaries, scattered strings, or external files detached from the notebook narrative.
Jupyter Config Cells keeps configuration next to the code that uses it:
# to: model_settingsabove YAML, JSON, JSONC, TOML, or ENV content- automatic config-language detection for obvious cells
- explicit cell actions for insert, set type, validate, export, and convert
- hidden runner cells that assign variables with Python standard libraries where possible
- no Python package, runtime extension, or
%load_extrequired
Example
# to: model_settings
model:
provider: deepseek
name: deepseek-chat
Run Convert to Variable, then use model_settings from later Python notebook cells.
Commands
Config Cells: Insert YAMLConfig Cells: Insert JSONConfig Cells: Insert JSONCConfig Cells: Insert TOMLConfig Cells: Insert ENVConfig Cells: Set Cell TypeConfig Cells: Validate CurrentConfig Cells: Export CurrentConfig Cells: Convert to Variable
Related Project
Jupyter Config Cells is published by Maplemx as an independent open-source project. Maplemx is also a core developer and maintainer of Agently, a framework for building production-grade AI applications with contract-first outputs, TriggerFlow orchestration, observable action calls, pause/resume/persist workflows, session memory, hierarchical settings, and pluggable action runtimes.
Development
npm install
npm run compile
Open this folder in VS Code and launch the extension development host.