How to Actually Build Agents with DeepSeek R1 in n8n - Without OpenRouter
Nate Herk | AI Automation via YouTube
AI Adoption - Drive Business Value and Organizational Impact
Gain a Splash of New Skills - Coursera+ Annual Just ₹7,999
Overview
Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn to integrate DeepSeek's R1 reasoning model directly into n8n workflows without using OpenRouter as an intermediary service. Discover how to leverage DeepSeek R1 as a cost-effective planning agent that can save up to 96% compared to OpenAI's o1 model while maintaining powerful AI capabilities for automation tasks. Explore the current limitations of DeepSeek R1, particularly its lack of built-in tool and function calling capabilities, and understand why it works best as a planning agent that develops step-by-step action plans for other AI models like GPT-4o or Claude 3.5 Sonnet to execute. Master the technical setup process for connecting DeepSeek directly through HTTP requests, eliminating the need for third-party routing services. Build a comprehensive two-agent system where DeepSeek R1 serves as the strategic planner, analyzing human requests and creating detailed execution plans, while a separate tools agent handles the actual implementation using traditional AI models with function calling capabilities. Access the complete workflow template and examine the specific prompt engineering techniques required to optimize DeepSeek R1's planning performance within n8n automation environments.
Syllabus
00:00 Quick Demo
00:46 Current Limitations of DeepSeek R1
02:18 Connect to DeepSeek without OpenRouter
04:31 Creating a DeepSeek Planning Agent
06:12 DeepSeek R1 Agent Prompt
07:07 Download This Workflow Free
07:53 Inspecting Tools Agent
09:32 DeepSeek R1 via HTTP Request
Taught by
Nate Herk | AI Automation