Відео

@@jamaslosubi sure, we have detailed guide in our community with all SQL queries

I just booked a demo and said that want to get access - they shared the link. I think they are still in Beta.

@@5_min_ai I see, thank you much for a response, I will try!

Hi, there is nothing special, I just use it to retrieve list of records from Supabase. These records contains JSON field and it breaks response (bug) for default nodes of Supabase and PostgreSQL in n8n.

For sure, you can find all workflows templates in our community for free. You will be able to copy and paste them into your n8n instance. Out first video was about selfhosting N8N with DigitalOcean - check this out to get your local version of N8N

@@mikehenderson-b5h you can find all templates for free in community and copy them. We have guide on a channel about self hosted N8N

Yeah, agree, in this video is too loud. Will finetune it in next video. But without music at all it looks boring, doesn't it? 🤔

@@5_min_ai with backgorund music it's bit chaotic for me, most of the creators who create similliar content do not use music in the background

It's Bubble.io My main stack for all projects - N8N, Bubble, Supabase

@@5_min_aido you connect supabase via api call to the bubble front end?

You're right, we quickly showed at the beginning of the video main path and then more focused on workflow itself. But basically - you can see how Airtable table looks like on 0:50 and google drive is simple folder that you can choose in first node. If you want to have call support - feel free to reach us in community and we will help you there.

Thank you, great point! Will make next part in dark mode

I used credentials for SQL connection. You can find them in Supabase. Open Tables - Connect button on top header - Connection String - Transaction Pooler. Also you can check detailed guide in our community - www.skool.com/5minai

Great catch Yes, it's mistake - it should be added in OpenAI payload under User content. Thx for this discovery!

@@Ali.Khaldi good point - new tools of AI agents have bug right now. When agent generates multiple items and try to run the same tool twice or more time this tool gets data from first item for all items. They told that it is fixed but it is still there. So use external workflows is more reliable solution for now

Thank you for tip! For sure, we are showing it for educational purposes and kill these key before video publishing.

Gemini model doesn't work with tools nodes in N8N - at least it didn't on 1.5, but looks like on 2.0 tools are supported. We will explore Gemini 2.0 and do overview of ability use it soon. Despite it has tools - it takes some time while n8n add new features.

@5_min_ai waiting for your upcoming tutorials and thank you very much for your time and support

Thank you for your time! Yeah, I tried enhancing feature in Premier Pro and got bad results. But I'm working on it 😊

@@5_min_ai Yeah it's unwatchable. May want to remake it?

@@ThatNerdChris hm, sorry for quality! I will check if still have raw record to remake it

@@llotusfn3354 hi You can find detailed guide and template itself in community It is simple API request where main param is query - the same you do in youtube search

thank you! 😊

nope, but you can do that with simple docker

@5_min_ai I know but there's problems in postgres host when using self hosted supabase. The url doesn't work, it ain't easy as the cloud one

No, you don't have to pin it. You can pin data in nodes for testing/development. I didn't get the last question - when AI agent triggers second scenario with switcher it send param "command" with one of the options from the list (name of tool like "search". Then in 2nd scenario switcher choose the path based on this param. Feel free to create post in community with other questions, I will help you with detailed screenshots.

Yeah, I figured it out. The issue was that my n8n version was outdated. It had a bug where, if the execute workflow was in the same workflow as the agent, I would get an error. Updating my n8n resolved the issue. I took a different approach but similar results. Your approach could be useful for my other projects. Thanks

Hi! How do you usually do this?

Cost is fine with gpt 4o-mini, a few cents. Gpt 4o is more expensive and has smaller context window. I use Sidekick browser - good for multiple separate sessions - I manage multiple project each of them has own bunch of accounts

@dirty_nocoder thanks a lot, very instructive. I'll check. Keep posting new video 👍👍👍

@ thank you!! Join our community to get more support in building agent Compare own

Sure, please create post in Community so we will be able to discuss it with screenshots there www.skool.com/5minai

Hi what step are you on? Feel free to join my community (it's free) and write DM (www.skool.com/5minai)

Sure! Added guide with scenario template and details

I don't know, I usually compare prices with alternative tools and it's x50 times cheaper. AWS and GC aren't user friendly in my opinion, especially for non-technical people.

Sure, I will create a video about it soon

You're very welcome!

@@chuckaspeer1 I made a fork from original one and added own scripts that automates installation

OpenAI assistants have their own memory within a thread. I don't know details about it, but in my opinion it is better than postgres memory - it only sends last X (e.g. 1000) tokens or last X (e.g. 5) messages included in prompt. It may interrupt the prompt itself and corrupt the results.

@@5_min_ai thank you for sharing this

@@5_min_ai Thanks a lot for the video... I'm looking for something like this but a model that is work in local using for example Ollama. Can be this possible please ? Maybe using the local PGVector in order to remember the important pesonal information while using personal assistant. Will be grate :-) Thanks again

@@MrDenisJoshua Sure! I haven't tried it but you just need to install Ollama on hosting and then use it in the scenario. Here is link on installation guide - github.com/n8n-io/self-hosted-ai-starter-kit And scenario example - n8n.io/workflows/2384-chat-with-local-llms-using-n8n-and-ollama/