n8n and tools
Automation with n8n
n8n is the backbone of many of my systems, including the AI agent newsroom. I build n8n workflows and agents end to end, and where n8n is not enough I add code, MCP servers and agents built with Claude Code and Cursor.
Why n8n, and when it is not enough
n8n is a workflow builder: trigger, steps, conditions, integrations. You can host it on your own server, so data never leaves for someone else's cloud and runs are not capped by a pricing plan.
For most business tasks that is enough: leads, CRM, notifications, autoposting, reports. AI nodes let you plug in a model and build a simple agent right inside a workflow.
But once there are many agents and you need memory, complex rules and a decision log, workflows turn into a tangle. Then I move the logic into a separate service and keep n8n for what it does best: integrations and delivery.
Tools I work with
- 01
n8n
Workflows, webhooks, AI nodes, self-hosted with no run limits.
- 02
Claude Code and Cursor
Vibe coding: quick scripts, scrapers, bots and websites built with AI.
- 03
MCP servers
Connecting agents to data: spreadsheets, CRM, keyword tools, databases, files.
- 04
OpenClaw
A personal agent on your own server: scheduled tasks, files, messengers.
- 05
Models
GPT, Claude, DeepSeek, Qwen, Grok: I pick the model for the task and the budget.
- 06
Custom code
Python and a database where speed and complex logic matter.
How I build a workflow
- 01
Review
What should happen, in which systems and with which exceptions.
- 02
Design
Steps, conditions, where AI decides and what to do on failure.
- 03
Build and test
The workflow on real data, with logs and retries.
- 04
Handover
Documentation and a walkthrough of how to change it yourself.
FAQ
Which is better, n8n or Make?
Make is convenient for getting started in the cloud. n8n can be self-hosted: your data stays with you, runs are not capped by a plan, and it is easier to write your own code inside a workflow. For systems that run a lot every day, I choose n8n.
Can you build an AI agent in n8n?
Yes. n8n has agent nodes: model, tools, memory. That is enough for one or two agents. For a team of agents with a decision log, the logic is better moved into a separate service.
Will you set up n8n on our server?
Yes. I install n8n on your server and configure access, backups and run history retention.
What is MCP and why does it matter?
MCP is a standard for connecting an AI agent to external services: spreadsheets, CRM, databases, files. Instead of a custom integration per task, the agent gets a clear set of tools with limited permissions.
Do you teach n8n?
Yes, one-on-one and for teams. See the AI agent training page for details.
Articles
n8n vs Make: Which to Choose for Business Automation
n8n compared with Make for automation: where data lives, how runs are billed, AI agents, custom code, learning curve. Who each suits and when to switch.
OpenClaw: What It Is and How to Install and Set Up Your AI Agent
OpenClaw is an open-source personal AI agent that lives on your computer or server and works through messengers. How to install it, configure it and keep strangers out.
What Is an MCP Server and How to Connect It to an AI Agent
MCP is the standard AI agents use to connect to spreadsheets, CRMs, databases and services. What an MCP server is, how it works, how to connect one and what to watch for security.
Related
A newsroom, a content pipeline, engineering data or an EnergyAI Ops pilot at your site. Tell me the problem - we'll work out where AI delivers and where it doesn't.