Hi n8n community! 
I recently published a Medium blog explaining how to optimize AI model settings (temperature, top-p, max tokens, etc.) to reduce hallucinations and build efficient n8n AI agents. This guide is perfect for:
- Developers working with AI in n8n workflows
- Teams struggling with unreliable AI outputs
- Anyone using OpenAI/Gemini nodes for automation
The blog breaks down complex parameters with simple examples and includes tested n8n configuration screenshots (like the OpenAI node settings). You’ll learn:
- How low temperature + top-p settings reduce “AI-made facts”
- Best practices for balancing creativity vs. accuracy
- Real-world use cases (customer support bots, email summarizers)
Check it out here:
https://medium.com/@ahzem/how-to-choose-the-right-ai-model-settings-e438ce7eddc4
I’d love to hear your thoughts:
- How do you tune AI settings in n8n?
- Have you faced challenges with hallucinations?
- Any tips to add for the community?
Let’s discuss it! 
*P.S. If you test these settings, share your workflow screenshots or results below! *
Why this works for the n8n community:
- Fits the “Built with n8n” category (showcasing practical AI workflow optimizations).
- Solves common pain points (hallucinations, unreliable outputs) highlighted in recent forum threads
- Encourages knowledge-sharing by asking for community feedback and experiences.
- Includes actionable takeaways with n8n-specific examples