n8n key links

n8n key links

Focusing on architecture....how it runs esp in the enterprise

A medium article: n8n Workflow Automation - Jin

From something called groove tech, another article

n8n deployment architecture diagrams

Search for: n8n deployment architecture diagrams

Note: from the first link here

n8n starter document, Nate Herk, a PDF

  1. General n8n docs are sparse
  2. See if this document throw some light
  1. n8n is a low-code,
  2. node-based workflow automation tool
  3. that allows users to connect multiple applications and automate tasks.
  4. It?s known for its visual aspect and flexibility
  5. it supports custom code and integrations with APIs.
  1. Zapier
  2. Make.com
  1. n8n is free to self-host, providing a cost-effective solution for those with technical expertise.
  2. For users preferring a managed service, n8n offers cloud plans starting at $20 per month, which include hosting and additional features like enterprise solutions!
  1. With free self-hosting
  2. and affordable cloud plans,
  3. n8n provides a budget-friendly solution,
  4. especially for businesses looking to scale their automation without incurring high costs.

An attempt to document from folks

  1. It is a forum
  2. it is an old thread that is archived
  3. to see if there are better learning guides...
  4. I don't think it is super helpful on first read....
  5. I will leave it here, in case if I have to come back
  1. Quote
  2. "However, as a person who doesn?t have that much coding experience, I was quite upset by the fact that n8n has a steep learning curve ? mainly due to the fact that there are no easy step-by-step guides on how to self-host n8n on the cloud or even local machine (though, I need a cloud-based solution because my laptop doesn?t always have the internet connection!) cheap and easy."

Summary of this to be later

Briefly, see if there are any take aways....

  1. AI Agent Node: Automates decision-making processes by utilizing AI models, Memory for context retention, and tools for task-specific actions.
  2. Summarization Chain: Condenses large datasets or text into concise summaries.
  3. Question and Answer Chain: Processes data to generate precise answers to user queries.
  4. Basic LLM Chain: Simplifies interaction with large language models for custom use cases.
  1. You can see these as various "horizontal general purpose" agents.
  2. Each with their general capability that can then be specialized in a work flow.
  3. In some way the work flow where they are composed can be seen as a vertical agent
  4. And each of these as general agents
  5. Likely this is a useful pattern for Enterprise adoption even outside of n8n
  6. For example the chat mode and agentic mode in ChpatGPT can be seen as 2 General agents
  7. Once you look at them that way you can separate them out into distributed reusable python backend APIs like the Chat API and Completions API
  1. Horizontal Agents as "pluggable" components in a larger application
  2. Be that work flows
  3. or applications
  4. As merely reusable components
  5. That might be drawback as well....

n8n attempt at documentation :) Mostly youtube videos

  1. Provides a summary of n8n in ONE place
  2. Points to a few YouTube Channels where one is likely to learn the rest
  3. Explains how the previously "work flow automation" (like a BPM tool) now integrates it with the AI agents, to be woven into that
  4. Lists the 4 key agents (nodes) that does this work
  5. It is a well framed document
  1. It is a well framed document summarizing how to get there, but fundamentally cannot solve the "YouTube" based documentation....
  2. Run time architecture is still a missing document, for I saw no explicit references
  3. It continues to to expose the challenges to enterprises where the installation, care, and scale etc are only sparsely touched upon
  1. Hope I am wrong
  2. I do like the self hosting model
  3. Expandability
  4. writing our own code as nodes
  5. However at its core it is a work flow system with AI thrown in
  6. It is not ground up UX system like OpenAI or Claude or MS Copilot
  7. Its UX is accordingly "diverse" and specialized for each work. Not uniform and horizontal, as it will likely be with the newer AI first platforms
  8. It does however gives a GREAT experimentation engine for "Creating IT companies" that have "strong development genes"
  9. Cost seems to be very encouraging for enterprises
  10. will conflict or overlaps with BPM approaches for work flows