Hey everyone, It's been a while since I was dreaming about a personal assistant (with very basic coding knowledge). and I discovered what truly was an AI agent and understood that it's something I can use to fulfil my wish. Though I have no budget. So I started to dig Youtube: - Self-host n8n: check - Using Ollama : check - Using Ngrok to go on the internet : check - fought with google Oauth 2 : I won And now my machine : an old macbook pro Intel. Ollama works and I find it already cool for conversation testing the different LLM's (Deepseek is giving me creepy vibe though with how it shows it's thought process). I'm using Nate's template for prompts foundation and I'm now reaching some limits: I'm managing to create events in my calendar and sometimes it's wrong in timing and duration. Though it can't delete or update my events and it just hallucinate the answer. My questions: - Is it a power problem because my computer is old? - Should I keep improving the prompt till It work or is there some trick by creating an agent for each calendar tool instead of giving all the tools to one agent? - Should I spend the money to get OpenAI or Anthropic solution to at least get some result and keep the motivation up and not give up on AI agent - Or buy a new computer adapted for my need? Does anyone have experience with getting a personal assistant working an intel processor i7 2,5GHz and AMD Radeon R9 M370X 2 Go? My thought is just get claude or chatGPT and then get yourself a VPS to get your first success with AI agent in general and go back to Open LLM's a bite later when you have power to run them Thank you very much for your advices, Byron PS: I placed it in general discussion because of how broad the topic seems. Let me know if I should place it in another category!

Hello everyone , hope you all are doing great . I am learning about n8n and get optimum thoertical knowledge of it . but when i went to apply this knowledge and create some projects i found almost everthing is paid but being a college student i cant effort it right now as i am just in learning phase , is there anyone who also faces this problem and stuck there ... and have the solution of it then please help me out. Thanks Waiting for your help. Saarthak

Hello, has anyone ever had the error 'The tool input received did not match the expected schema' on the AI ​​Agent node? Lately I've been getting this error frequently whenever I need to process a tool, using an OpenAI model (through OpenAI node and OpenRouter) and with Gemini Flash. Does anyone have any explanation or how to solve this? Is it a problem with LangChain itself?

Let's get to know each other! Comment below sharing where you are in the world, a career goal you have, and something you like to do for fun. 😊