Creo que Python puede desbloquear usos potenciales y facilitar el desarrollo de flujos de trabajo para programadores nuevos.
Comencé a programar usando Python y me dediqué a automatizar procesos que principalmente involucraban datos tabulares (bases de datos, archivos de Excel o CSV, Google Sheets y PDFs)
Si pudiéramos usar Python dentro de n8n, al igual que en un IDE (me refiero a instalaciones fáciles e importaciones de paquetes y módulos), podríamos simplificar muchas automatizaciones y ayudar a un segmento de usuarios a entender mejor y más rápidamente los nodos de código (para quienes no son competentes en JavaScript).
Además de entender mejor el código más rápidamente, sería genial probar todos los diferentes tipos de uso de paquetes y módulos. El potencial podría ser ilimitado (pandas, polars, uso de IA estrecha determinista y muchos otros paquetes prácticos)
Pandas, Numpy y Datetime son mis paquetes principales para el procesamiento de datos tabulares.
@Sebastian_Arias You’re not starting from zero — Code nodes already run Python, and on self-hosted you can get much closer to “IDE-like” packages via task runners.
What works today:
Code node → Language: Python for small transforms (Pyodide / runner depending on version). Fine for datetime and light logic; not a full pandas IDE on Cloud.
Self-hosted task runners for real packages (pandas, numpy, etc.): enable external runners and add extras in the runner image / documented “extra dependencies” path (see Task runners docs). That is the supported way to import pandas without hacking the main n8n container.
Heavy tabular work many people keep outside the Code node: CSV/Sheets → HTTP Request calling a small FastAPI/Flask with pandas, or a dedicated Python container. Keeps n8n as orchestration.
Minimal pattern that scales better than stuffing everything into one Code node:
Then HTTP Request from n8n. On Cloud you will not get arbitrary pip install inside Code; vote the feature and use a sidecar or Sheets/DB nodes for tabular work until runners are available on your plan.
If you say Cloud vs self-hosted + version, I can narrow to the exact Code vs runner path for your setup.
What I want to promote is the idea of a “Plug and play” method for Python-first users. Having workarounds outside n8n when coding, gives the process a lot of friction.
To be clear, I do use n8n as orchestration tool as well on many occasions, but it would be great to be able to simply do it all in there. The whole system becomes less coupled, and you don’t have to jump from app to app and using complex directories for executing specific files of code.