Sui Serverless MCP Core

What is Sui Serverless MCP Core?

This server implements the backend infrastructure for enabling Model Context Protocol (MCP) services on the Sui blockchain using zkLogin. It is part of the Sui Butler system and is built with the AWS Serverless Stack, providing APIs, databases, and a user dashboard for managing access keys and approving transactions. It is designed for developers building AI‑powered tools that interact with the Sui blockchain while leveraging zkLogin authentication.

How to use Sui Serverless MCP Core?

Deploy the full‑stack application to AWS using the AWS Amplify infrastructure code. After deployment, users authenticate via zkLogin (only Google accounts supported in this version) to access the dashboard. Transaction requests from an MCP‑compatible AI client are stored in the database with a pending status; the user then visits the dashboard to manually approve them. Alternatively, the private key mode (documented in the main repo) can be used without this backend.

Key features of Sui Serverless MCP Core

• Serverless architecture built with AWS Amplify and Next.js
• zkLogin authentication via Enoki (Google accounts only)
• Dashboard for managing access keys and approving transactions
• Stores metadata associated with zkLogin accounts (email and wallet address)
• Transaction approval workflow with pending status and manual confirmation

Use cases of Sui Serverless MCP Core

• Enabling MCP services on Sui with zkLogin for AI clients
• Managing and approving Sui blockchain transactions via a web dashboard
• Building serverless backends that combine blockchain interactions with authentication
• Providing a secure way for users to approve pending transaction requests

FAQ from Sui Serverless MCP Core

What is the difference between using this backend and private key mode?

This backend uses zkLogin for authentication and a dashboard for manual transaction approval. Private key mode, detailed in the main repo, allows use without the backend.

What runtime dependencies does the server require?

It relies on AWS Amplify (serverless stack), Next.js for the frontend and API routes, Enoki for zkLogin, and a database for metadata storage.

Where is user data stored?

Metadata associated with zkLogin accounts – specifically each user’s email and wallet address – is stored in the database.

What are the known limits of this version?

The zkLogin authentication currently supports only Google accounts.

How does authentication and transport work for transactions?

Users authenticate via zkLogin through Enoki. Transaction requests are pushed from an MCP‑compatible AI client to the backend, stored as pending, and then approved manually on the dashboard.