Day 57 of 100 Days Agentic Engineer Challenge: Prototyping with n8n
I’ve noticed that I often struggle to come up with agent ideas when coding. It takes me a lot of time to plan workflows before I can start automating things. I also run into issues with OAuth 2.0 for social media and finding a good long-term memory solution.
To make things easier, I’ve decided to prototype every idea with n8n first. This low-code tool will help me manage workflows, long-term memory, and even experiment with self-learning features. I’ll build my ideas in n8n first, and later, I might explore an agentic framework or extend an existing n8n workflow with some custom backend code.
We’ll start by setting up n8n on a self-hosted cloud solution. Then, we’ll create a simple chatbot using the n8n Agent node and Window Buffer memory. But first, let’s take a step back and review my daily task routine.
Daily Tasks Routine
💪 Physical Activity — I did 80 squats and 40 crunches.
😴 Seven Hours of Sleep —I slept for 7 hours, went to bed earlier, and woke up before 6 AM. I like this much better than staying up late at night.
🤖 AI Agent — I’m building a true AI agent.
⏳ PAIC — In queue.
📊 Data Science — In queue.
If you want to know what all these tasks are about, read the introduction to the 100 Days Agentic Engineer Challenge.
What is n8n?
n8n is a node-based, event-driven workflow automation platform leveraging a distributed architecture for orchestrating API integrations, data transformations, and system interactions through visual pipelines. Built on TypeScript with JSON-LD schema definitions, it supports 400+ connectors via REST/SOAP endpoints while enabling custom JavaScript/Python nodes for bespoke logic and ETL pipelines. The platform offers self-hosted deployment via Docker/Kubernetes with OAuth2 authentication flows, JWT token validation, and end-to-end encryption for enterprise-grade security compliance (SOC2, GDPR-ready). Its execution engine features parallel processing with worker queues, atomic transaction rollbacks, and distributed tracing through OpenTelemetry integration. Designed for technical teams, n8n provides CLI tooling, Swagger API documentation generation, and Webhook signatures for building complex automation topologies across hybrid cloud environments.
Self-hosted n8n solution in the Cloud
Yesterday, I shared a great resource for getting free credits on various AI tools. One of them was Render, a cloud hosting and scaling solution for apps.
You can use Render to self-host n8n — just search for the n8n Docker image. Here’s what I found: docker.n8n.io/n8nio/n8n.
To set it up, go to your Render dashboard and add a new web service. Before clicking Connect, make sure to configure a few Environment Variables first.
Important! If you want to use tools like Google Calendar, Gmail, and others, you’ll need an OAuth Redirect URL. By default, it’s set to localhost, which won’t work.
To fix this, set N8N_EDITOR_BASE_URL and WEBHOOK_URL to your full URL, including https://. If you deployed the service before setting this up, don’t worry—you can update it afterward and simply restart the container.
Here’s the link to all environment variables:
🔗 n8n Environment Variables
Now, you can click “Connect”. Once deployed, go to the Render-provided link, add the owner account, and you’re good to go!
Start with a basic workflow, and if you want to connect with the n8n API, check out the docs here:[your_full_url]/api/v1/docs/
I just created my first n8n automation — it checks for events in my Google Calendar and adds new ones automatically. Of course, this simple automation isn’t a full agent, but it does include an Agent node, which does a great job of choosing the right tool for the task.
Next, I’ll add Telegram to trigger the workflow and gather feedback. After that, I’ll work on integrating long-term memory.
