Introducing the Suga MCP Server

The Suga CLI now includes a Model Context Protocol (MCP) server. This gives AI coding assistants like Claude, Cursor, and other MCP-compatible tools direct access to your Suga infrastructure and complete documentation.

What is MCP?

The Model Context Protocol is a standardized way for AI assistants to access external systems and data sources. Instead of relying solely on training data, AI assistants can query live systems and documentation through MCP servers.

The Suga MCP Server implements this protocol to provide two types of access:

• Documentation search, providing access to the full Suga Docs
• Infrastructure API access, allowing agents to perform most Suga task, like designing applications, platforms, plugins and more

How It Works

The MCP server is implemented as a new command in the Suga CLI (`suga mcp`). It runs as a process that communicates with AI assistants over stdio using the Model Context Protocol specification.

Under the hood, the server registers two types of capabilities:

Tools: Functions that AI assistants can call to perform operations. These include listing platforms, fetching templates, retrieving plugin manifests, and generating Terraform from `suga.yaml` files. Each tool is backed by calls to Suga's existing APIs, authenticated using your `suga login` credentials.

Resources: Documentation that AI assistants can search and retrieve. The server acts as a proxy to Suga's documentation, allowing AI assistants to query the latest information without relying on potentially outdated training data.

The MCP server is built into the CLI rather than exposed as a web API. This design decision was made because many of Suga's features are local-first and only available through the CLI—like generating Terraform from local `suga.yaml` files. The CLI can also access all web API functionality, so implementing the MCP server in the CLI provides the most complete feature set.

Authentication is handled through the same mechanism as the rest of the CLI. The MCP server runs with your credentials and has access to exactly what you have access to—no more, no less. It doesn't have special server privileges; it simply acts as you.

Documentation Access

The MCP server can search and retrieve information from the complete Suga documentation library. When you ask your AI assistant a question about Suga, it searches the docs in real time rather than relying on potentially outdated training data.

Examples:

• "What persistence resources are available in Suga?"
• "What's the syntax for environment-specific configuration?"
• "Show me how to give my service `read` access to an S3 bucket"

The AI searches the current documentation and provides answers based on the latest information.

Infrastructure API Access

The MCP server exposes several tools to interact with Suga infrastructure, including:

• Platform discovery to list and inspect available platforms, revisions, and configurations
• Template exploration to browse application templates with full specifications
• Plugin library access for up to date plugin manifests and schemas
• Build APIs to generate Terraform from `suga.yaml` configuration files

This means your AI assistant can discover what platforms your team has access to, what templates are available, and what plugins each platform provides. It can use this information to provide infrastructure advice that's specific to your environment and application.

How to Use Suga MCP

When you ask your AI assistant to add a PostgreSQL database to your application, it can:

1. Search the documentation for database configuration syntax
2. Discover which platforms you have access to
3. Check what database plugins are available on those platforms
4. Generate the appropriate `suga.yaml` configuration

The result is infrastructure code that matches your specific environment rather than a generic example.

Adding a Database to Your Application

Before: Manually search docs, copy-paste YAML, configure and test.

After: Ask your AI assistant "Add a PostgreSQL database to my app." It searches the docs, discovers your available platforms, and generates the correct `suga.yaml` configuration automatically.

Exploring New Platforms

Before: Manually browse platform specs, compare options, try to understand which plugins are included.

After: Ask "What platforms support serverless containers?" Your AI assistant lists available platforms, explains their differences, and shows you exactly what resources each provides.

Migrating Infrastructure

Before: Read migration guides, update configurations manually, test and iterate until any breaking changes are resolved.

After: Ask "How do I migrate from platform version 1 to 2?" Your AI assistant compares platform manifests, identifies changes, and guides you through the update process.

Access Control

The MCP server respects your team's access controls. AI assistants can only see platforms, templates, and resources that your team has access to. Public resources are available to everyone, while private team resources remain restricted.

Setup

Setting up the Suga MCP Server requires updating to the latest version of the Suga CLI and following the MCP Integration Guide to connect the server to your AI tool of choice.

The integration works with Claude Code, Cursor, and other MCP-compatible AI tools.

What's Next

Future updates will include:

• Tooling to improve platform creation and maintenance using agents
• Agent powered plugin development, allowing you to build new Suga Terraform plugins with assistance
• Enhanced code generation with validation

The code is open source. If you run into issues or have ideas for new capabilities, open an issue or PR on GitHub.

Availability

The Suga MCP Server is available now in the latest version of the Suga CLI. For complete setup instructions, see the MCP Integration Guide.