The Ultimate FREE n8n + MCP Setup Guide (No-Code Tutorial)

Key Concepts

Docker, mCP (Meta Control Protocol), nadn, Community Nodes, AI Agents, Environmental Variables, Containerization, Localhost, API Keys, Tool Usage, Workflows, Credentials, Firecrawl, Brave Search, AI Model (OpenAI, Ollama), Community Node Packages.

Docker Installation and Setup

1. Downloading Docker Desktop: The video begins by instructing users to download Docker Desktop from docker.com, selecting the appropriate version for their operating system (Apple Silicon, Intel Mac, or Windows).
2. Installing Docker: After downloading, the user is guided to install Docker Desktop by dragging the Docker image to the applications folder.
3. Opening Docker: Once installed, the user opens Docker Desktop, which displays containers, images, and volumes. The video emphasizes that extensive knowledge of these elements isn't necessary for the tutorial's purpose.
4. Pulling the nadn Image: The user searches for "naden" within Docker Desktop and pulls the official naden.io/naden image, which has over 100 million pulls.
5. Running the nadn Container: After the image is pulled, the user clicks the "Run" button to create a new container from the nadn image.

Configuring the nadn Container

1. Naming the Container: The user is instructed to give the container a descriptive name (e.g., "Neden container").
2. Setting Host Port: The host port is set to 5678. This port will be used to access nadn in the local environment.
3. Volume Configuration:
   • Host Path: The user creates or selects a local folder on their machine (e.g., "naden Docker" in the Documents folder) to store nadn data. This path is specified in the Docker container settings.
   • Container Path: The container path is set to /home/node/data, which is the location within the Docker container where nadn expects to store its data.
4. Environmental Variable:
   • A crucial environmental variable is set: N8N_COMMUNITY_NODE_ALLOW_TOOL_USAGE is set to true. This enables the use of community nodes, specifically the mCP node, within nadn.

Accessing nadn and Setting Up Account

1. Accessing Localhost: After running the container, nadn becomes accessible via localhost:5678 in a web browser.
2. Setting Up Owner Account: The user is prompted to create an owner account for their nadn instance.

Installing the mCP Community Node

1. Accessing Settings: Within the nadn interface, the user navigates to "Settings" and then "Community Nodes."
2. Installing the Node: The user installs the @n8n-community-nodes/n8n-nodes-mcp community node package.
3. Verifying Installation: After installation, the user verifies that the mCP client node is available in the workflow editor by searching for "mCP." If it doesn't appear, refreshing the page is recommended.

Setting Up mCP Credentials and Tools

1. AI Agent Setup: The user adds an AI Agent node to their workflow and selects a chat model (OpenAI or a local model like Ollama).
2. Adding mCP Client Tools: The user adds two mCP client tools to the AI Agent: one for listing tools and one for executing tools.
3. Firecrawl Credentials:
   • Command: npx
   • Arguments: -y firecrawl mcp
   • Environmental Variable: FIRECRAWL_API_KEY=<your_api_key>
   • The user obtains a Firecrawl API key by creating a free account on the Firecrawl website.
4. Brave Search Credentials:
   • Command: npx
   • Arguments: -y --model context_protocol_server brave_search
   • Environmental Variable: BRAVE_SEARCH_API_KEY=<your_api_key>
   • The user obtains a Brave Search API key by creating a free account on the Brave Search API website.
5. Configuring the List Tools Node: The mCP client tool is configured to "List Tools."
6. Configuring the Execute Tool Node:
   • The mCP client tool is configured to "Execute Tool."
   • The tool name is dynamically grabbed using the expression {{ $json["data"]["tool_name"] }}.
   • Tool parameters are set to be defined automatically by the model.

Testing the Workflow

1. Testing Firecrawl: The user tests the Firecrawl integration by asking the AI Agent "what tools do you have." The AI Agent lists the available Firecrawl tools. The user then asks the AI Agent to "scrape the info from aiworkshop.me," which triggers the Firecrawl execute tool.
2. Testing Brave Search: The user tests the Brave Search integration by asking the AI Agent "what tools do you have." The AI Agent lists the available Brave Search tools. The user then asks "what are the best sushi restaurants in San Francisco," which triggers the Brave Search execute tool.

Conclusion

The video provides a step-by-step guide to installing and configuring nadn with mCP support using Docker Desktop. By leveraging Docker, users can run nadn and its associated services locally and for free. The tutorial covers essential configurations, including setting environmental variables, configuring volumes, and installing community nodes. It also demonstrates how to integrate and test mCP tools like Firecrawl and Brave Search within nadn workflows, enabling AI-powered automation and data extraction. The video emphasizes the importance of the community and provides links to resources and support channels for further assistance. The key takeaway is that users can create powerful AI-driven workflows locally without relying on cloud-based services, offering greater control and privacy.