When Web3 meets MCP: Apillon MCP Server now live on GitHub!

Apillon

May 12, 2025 • 3 min read

No time for docs? Pair the Apillon MCP Server with an LLM and deploy faster.

Ever wondered what life would have been like if you didn’t have to read any documentation, but could still use AI to help you deliver the feature you have always wanted?

While LLM chat interfaces have surely unlocked tremendous value in the economy, the AI has rapidly been shifting its attention to MCP (Model Context Protocol tooling), which unlocks a standardised way to define and allow LLMs to access data sources through a set of functions, commonly referred to as tools.

Apillon goes MCP

Our development team decided to tackle this side quest head-on recently, and we are happy to announce that you can now use Apillon’s hosted MCP server. This allows you to hook into the functions of the Apillon platform—namely, the functions tied to some of our most popular services—Web3 hosting, storage, and NFTs.

Surely this sounds like rocket science and we can hear you thinking it must be a pain in the butt to hook up to your tooling. Think again, as all you need is Node.js installed locally on your machine and a simple general configuration, such as the one provided below.

{
    "mcpServers": {
      "apillon-mcp-server": {
        "command": "npx",
        "args": ["-y", "@apillon/mcp-server"],
        "env": {
          "APILLON_API_KEY": "<APILLON_API_KEY>",
          "APILLON_API_SECRET": "<APILLON_API_SECRET>"
        }
      },
      "filesystem": {
        "command": "npx",
        "args": [
          "-y",
          "@modelcontextprotocol/server-filesystem",
          "/Users/<your-username>/Desktop"
        ]
      }
    }
  }

If we break it down in even simpler terms, the config above defines two separate MCP server namespaces (apillon-mcp-server and filesystem):

The first one is maintained by Apillon.
The second, more general one, gives LLM access to a specific folder on your local filesystem.

To state it plainly, this means that you can now ask your LLMs to help you upload a static website from a subfolder on your desktop to Apillon. Provided you have filled in the API keys correctly (and those keys have permission to manage the Apillon hosting and storage service in your account).

The LLM will simply make the appropriate calls to the Apillon API and make sure your wish is its command.

Worried you might not have given it all the required information? That’s the best part. MCP definition allow for input validation and the LLMs are often smart enough to figure out which information you have provided, which information you still have to fill in and in case of an API error often find ways to self-correct, so that they can complete the task at hand as effortlessly as you can say AI or Web3.

Stop talking, where is it?

For those of you who can’t wait any longer, you can head over to our GitHub repository and test it out yourself. A more in-depth, step-by-step guide is coming shortly (a.k.a. in a couple days) in multiple formats, so that you can follow along and finally make that important first step on your Web3 journey.

Have fun and keep building.

⧓ About Apillon

The Apillon platform serves as a unified gateway to the Web3 services provided by linked Polkadot parachains. Following the multi-chain vision, Apillon powers the transition of developers to Web3, simplifying its adoption in the real economy, and expanding its versatility as the ecosystem grows. With Apillon, Web3 services are within reach for every developer, regardless of their background and experience with blockchain technology.