DEV Community

Cover image for Day 10 of 60 of learning AI automation in public
Cess Mbugua
Cess Mbugua

Posted on

Day 10 of 60 of learning AI automation in public

Today I solved a problem that has been sitting in the back of my mind since Day 1.
My n8n workflows and my Python scripts have been living in completely separate worlds. Today I connected them.


## The Problem

I've built powerful automation pipelines in n8n. I've built Python scripts that call Claude directly. But they couldn't talk to each other.

Every time I needed complex data transformation or custom AI logic that n8n couldn't handle natively. I had to build a whole separate system with no connection between the two. Duplicated effort. Disconnected systems. Limited capability.

This is the gap that stops most automation engineers from reaching the next level.


## What I Built

A FastAPI Python server that runs locally and exposes Python logic as HTTP endpoints. My n8n workflows send data to these endpoints via HTTP Request nodes, the same way they call any external API.

  • Python processes the data
  • Calls Claude API
  • Returns a structured JSON response back to n8n

From n8n's perspective, Python is just another API to call.
From Python's perspective, n8n is just a client sending requests.
Both systems stay independent but now work together seamlessly.


## New Skills This Introduced

FastAPI: building a Python web server
@app.post() decorator: creating API endpoints
Pydantic BaseModel — validating incoming request data
uvicorn: running the FastAPI server
HTTP Request node in n8n: calling Python from a workflow
async def: asynchronous function basics


## What I Learned

The biggest mindset shift today was realising that every tool communicates through HTTP. n8n, Make.com, Zapier, they all have HTTP Request nodes. That means any Python logic I wrap in a FastAPI endpoint becomes instantly callable from any automation platform in the world.

I didn't just learn FastAPI today. I learned how to make Python a first-class citizen in any automation stack.


🔗 Full project on GitHub → https://github.com/mbuguacessy-glitch/ai-webhook-processor

Day 10 of my 60-day public AI automation build challenge. Learning by doing.. every day.

50 more to go.

python #fastapi #n8n #claudeapi #automation #learninpublic #buildinpublic #100DaysOfCode

Top comments (0)