Bart Slodyczka

n8n hack #207

If you need to return the entire data object from a previous node (even if it is 5 nodes back) use the below expression:

return $('NodeA').all();

Where 'NodeA' is the name of your target node.

This will pass through the object EXACTLY as is.

Use case: I have a flow where I need to fan out the data, process it, and then merge it back, to join up to two other paths. At the merge point, I need to original data object, in its entirety.

Skip the Carousel

1 2 3

2 days ago | [YT] | 91

Bart Slodyczka

Video on AI Jailbreaking coming soon (How To Defend Against It)

This is using gpt-5.1, which is better than the previous gpt models. Ie this is alarming haha.

Granted my prompt is simple, and the workflow is basic. But I think this is how a lot of people are building agents (right at the fringe of the interaction, without buffers).

So I will make an intro-level video to try bring some more awareness to this.

There are ppl in my Skool community who are deep in the weeds on AI security. If you want to improve in this area, come and start convos.


My community is free: www.skool.com/flow-state

Skip the Carousel

1 2 3

3 days ago | [YT] | 66

Bart Slodyczka

I just got a Raspberry Pi 🥹 😁

I am so excited! I've recently been super interested in the stuff 'below the code', pretty much starting at the level of 'how does electricity, that runs through circuits, actually run code'

I was initially planning on get a breadboard kit (off an old school Youtuber's website, his name is Ben Eater) so that I could literally start from scratch

But I decided to get the Raspberry Pi. This is essentially a mini-computer that you can use to do things like:

- Add a camera and you have vision
- Add a sensor and you can do motion sensing
- Add a local AI, a Mic, and a Speaker, and you can chat with AI
- Add a screen and use as a computer
- A thing called a 'Pi Hole' which acts as an add blocker
- INSTALL n8n ON IT! haha

I'm sure that list doesn't do it justice, and in truth, I am a beginner with this. So I am sure I will learn A LOT by using this thing. I want to try write mini programs for it. I want to hook up cameras, etc.

Super super excited for this!!

Have you ever heard of a Raspberry Pi, or have you used one before? Please share any cool tips and tricks. I may also end up recording mini videos about this thing, if you guys wanna see 💪

Skip the Carousel

1 2 3 4 5

2 weeks ago | [YT] | 74

Bart Slodyczka

Hey guys! Just finished recording a video on Relational Databases. To be used as a introductory course for beginners AKA new age AI devs who are using AI tools to build apps and automations, but may need a couple pointers about the developer world.

Plenty more depth that can come out of topics like this. Databases are a massive unlock if you are serious about building things with AI. Will have the video edited and released in the coming days. Stay tuned and hope you enjoy :)

Thanks, Bart 😊

Skip the Carousel

1 2

1 month ago | [YT] | 64

Bart Slodyczka

OpenAI Agent Builder did not kill n8n

I don't even think OpenAI would say they are trying to compete with n8n. Their platform is more AI-centric, whereas n8n is literally the heart and soul of automation.

Great Stack AI article link here: www.stack-ai.com/blog/stackai-vs-openai-agentkit

I used a screen shot of the table from stack AI for the last 2 images in this post. But there is a ton of other info on that article that is worth reading :)

Great comparison.

I think the ppl who think OpenAI kills n8n are not using n8n correctly, or for the right reasons. I don't buy into hype either. Everyone is just trying to coexist and build cool shit together. Stay focused on your goals and keep your eyes on the prize.

Skip the Carousel

1 2 3 4 5

1 month ago | [YT] | 90

Bart Slodyczka

Flick through the images and you will see that the Anthropic chat node has a setting where you can 'Enable Thinking'

When you turn this on your agent actually spends tokens 'thinking' about your question, just like what you get in the Claude chat interface.

The last image shows how the output response contains two objects:

• One is the type: thinking - the thinking tokens
• The other is type: text - the actual response

The OpenAI chat node also has a similar setting for 'Reasoning Effort' (the thinking equivalent) where you can set it to Low, Medium, and High, for certain models

There is a bit of nuance to the thinking mode:

• This is different to using the Think Tool
• You cannot get access to the extended thinking mode if using OpenRouter chat node
• In the Anthropic Dev Docs, they speak about how you can use the extended thinking mode to see how the model processes your prompt (link below)
• In the same Docs, Anthropic says if you want to give the model <1k tokens to 'think', you should instead use CoT (chain of thought) prompting using <thinking> XML tags (link below)

I regularly go through the n8n updates page to see what new stuff comes out. You should too!

I also made this video going into a bit more depth about what was mentioned in this post

Video link: https://youtu.be/eEIdHWrsaNc
Anthropic link 1: docs.claude.com/en/docs/build-with-claude/prompt-e…
Anthropic link 2: docs.claude.com/en/docs/build-with-claude/prompt-e…

Skip the Carousel

1 2 3 4

1 month ago | [YT] | 50

Bart Slodyczka

Did you know Claude has memory, similar to ChatGPT?

Go to Settings > Capabilities > Toggle On

Open a new chat and ask!

Searching past chats is available to all users on paid plans (Pro, Max, Team, and Enterprise plans) on the web, Claude Desktop, and Claude Mobile apps.

Read more: support.claude.com/en/articles/11817273-using-clau…

Skip the Carousel

1 2

1 month ago (edited) | [YT] | 23

Bart Slodyczka

If you use n8n, don't migrate your production databases to Data Tables (... yet)

This is n8n's own warning from their implementation code:

const DATA_TABLE_WARNING_MESSAGE = `[Data table] The Data table module is experimental and subject to change.
Any tables added before the official release may become inaccessible at any point. Use at your own risk.`;

Which makes sense cos the release is still in Beta. But I have heard and been asked by a few ppl about migration.

I'll be making an 'everything you need to know' video soon :)

Gunna go over the look at the docs, the deep wiki, have already tried to use the API (but don't think it's publicly facing lol), built a mini full stack app (using form node, logic nodes, and data tables), and will also chat about some limitations in this current form. All this to say, give it 3 months and I think this thing will be 10x as useful.

very excited! 💪

1 month ago | [YT] | 53

Bart Slodyczka

Howdy legends!

You may have seen I am creating a series around production ready n8n systems.

So far I have created videos about:

1. Building workflows using Environments: this is a pretty interesting on how to approach building out automations. Eg how to purposely try break things, then how to test in bulk, and finally how to deploy live

2. How to Scale self-hosted n8n: I run through scaling the server && going from 'main' to 'queue' mode, I test everything using n8n's benchmarking tool, and compare results to their cloud version. The outcome is that we scaled from 20 executions/ second to 90 executions/ second

3. Webhook Security: which also covers correct server set up when self hosting, HMAC, and payload validation. Some of this stuff I have learned from other much more experienced people

4. Overview of what production ready means: this is a holistic overview of production ready, and shows you how things like logging, databases, webhook security, managing multiple clients, literally all forms different parts of the production ready web

The next video will be about databases and logging workflow events. I've been interested in databases for a long time and am excited to create the next video.

If you haven't watched one of these videos yet, go and check them out. My favourite is the 4th video, the overview of what production really means. Give you a nice overview of how everything connects together.

I created all these videos because people in my skool community are asking about them. I'm seeing more and more a demand for this type of developer knowledge. Which is actually difficult to find. Each video takes me 10-20 hours to learn and understand. My goal is to create a series of about 5-10 videos so that anyone can build like a pro.

Hope you are enjoying this series!

Skip the Carousel

1 2 3 4 5 6 7

1 month ago | [YT] | 166

Bart Slodyczka

I just came across an open source repo on github that lets you build voice agents

Repo: github.com/pipecat-ai
Dev docs: docs.pipecat.ai/getting-started/introduction

Pipecat is a python based framework for building voice (and multimodal) agents
You can self host this, like n8n, and wire in your own LLMs
You could use this instead of Retell AI or Vapi

→ Downsides:
- Need to know how to deploy and manage
- Latency might be slower cos the engineers at Retell, etc. would have built smooth pipelines
- Less 'out of the box' features

→ Upside:
- Cheaper at scale
- More LLM vendors
- Multimodal experience, eg can deploy a video chat agent on your website

→ Repo explained:
- Core, the engine that 'powers' everything: github.com/pipecat-ai/pipecat
- Flows, visual conversation editor: github.com/pipecat-ai/pipecat-flows
- UI Kit, internal and external UI kits: github.com/pipecat-ai/voice-ui-kit

Might make this a September project in my Skool: www.skool.com/flow-state
Join if you want to vote and follow the convo

2 months ago | [YT] | 84