n8n AI Agent for DVM MCP

What is n8n AI Agent for DVM MCP?

n8n AI Agent for DVM MCP is a set of n8n workflows that create an AI Agent capable of discovering and using Model Context Protocol (MCP) tools served as Data Vending Machines (DVMs) over the Nostr network. It is designed for developers and AI enthusiasts who want their LLM to find and interact with tools that are not locally installed.

How to use n8n AI Agent for DVM MCP?

Install a self‑hosted n8n instance and add the Nostrobots Community Nodes. Import the five provided workflows from GitHub using the “Import from URL” feature. Configure credentials for an OpenAI API (or other LLM), SerpAPI, a Nostr private key, and a Postgresql/Supabase database. Set workflow variables for assistant name, assistant npub, Nostr relays, and user localisation. The main workflow “DVMCP Agent Example v1.0” must be connected to its four tool subworkflows; on first install you may need to manually reconnect them.

Key features of n8n AI Agent for DVM MCP

• Discovers MCP tools by querying the Nostr network as DVMs.
• Posts requests to discovered tools and reads responses over the network.
• Works without installing each MCP server locally on the same system.
• Uses a low‑code platform (n8n) for easy workflow creation and modification.
• Integrates the DVMCP bridge to connect MCP servers with Nostr DVMs.

Use cases of n8n AI Agent for DVM MCP

• An LLM that can find and call a calculator, weather, or translation tool it does not have installed.
• Automating multi‑step data‑vending tasks where tools are scattered across different MCP servers.
• Demonstrating how AI agents can perform tool discovery and execution over a decentralized network.

FAQ from n8n AI Agent for DVM MCP

What is DVMCP?

DVMCP (Data Vending Machine Context Protocol) bridges Model Context Protocol servers with Nostr’s Data Vending Machine ecosystem. It allows any MCP server to serve tools over the network and any agent to access all tools served.

What prerequisites are required to run this agent?

You need a self‑hosted n8n installation, the Nostrobots Community Nodes, credentials for an LLM (e.g., OpenAI API), SerpAPI, a Nostr private key, and a Postgresql/Supabase database.

How does the agent discover tools?

The agent’s system prompt instructs it to query the Nostr network for available DVM MCP tools when a user request requires an uninstalled tool. It then posts requests, waits, and reads responses.

Where does data live during operation?

Credentials and workflow variables are stored in n8n. Tool requests and responses are broadcast over the Nostr network. The agent awaits and reads responses from the same network.

Are there any known limits or required transports?

The agent relies on the Nostr network for communication and waits up to 45 seconds for tool responses. It uses the DVMCP protocol and requires a Nostr private key for signing events. No other transports are documented.