ElevenLabs is constantly improving its platform. I used this to create a voice agent that connects directly to the MCP server in N8N. This eliminates the need for any configuration within N8N itself—depending on who's doing what, of course.

You can now connect multiple LLMs (OpenAI, Claude, Gemini, DeepSeek etc) and dynamically pick which model gets used based on your input data. This lets you build agentic routing: - If the topic is legal → use Claude - If the cheapest option is needed → use DeepSeek - If the input is Estonian → route to GPT-4 *How to use* Video tutorial coming...

Been working on a full breakdown of how RAG Agents actually use vector databases in real workflows and it all works inside n8n. This module is dropping soon inside the community. Keep grinding.. - Rico

I just tested Exa.ai and its pretty insane how good it is at finding the right info fast. Way better at deep research than most LLMs (like ChatGPT etc..) but slightly better than Perplexitys. Perfect for n8n agents doing: 🔍 Web search 📚 Deep research 📎 Real source citations

Hey guys i just find this crazy thing - String 👇 🔗 string.com - it's like Lovable or bolt.new, BUT you can literally build & deploy AI Agents with one single prompt (I will leave screenshots below (tried to build a auto email answer with labeling and it built it for me, WITH ONLY ONE PROMPT, na this is crazy)) You just have to connect all you accounts and it works. You will get limited credits/tokens with free version, same like Lovable 🎥 I’ll drop a quick video demo under this post soon 👇 So let me know what you want to see.. -Rico