Firefly MCP Server is a TypeScript-based MCP server designed to seamlessly integrate with the Firefly platform, acting as a bridge between AI models and your cloud and SaaS resources. It empowers AI to discover and codify resources across your connected accounts, converting them into Infrastructure as Code. Utilizing secure authentication with access keys, Firefly MCP simplifies resource management by enabling AI to find any resource and represent it in a codified format.

This server enhances AI model functionality by providing real-time access to cloud infrastructure data, solving the developer pain point of manually managing and translating cloud resources. It integrates smoothly with AI tools like Claude and Cursor, allowing natural language queries to translate into infrastructure code. Firefly MCP can be integrated via standard input/output or HTTP/SSE, offering flexibility in how AI models interact with your cloud environment.

Resource Discovery Across Clouds

Firefly MCP enables AI models to discover and access resources across various cloud and SaaS environments. It acts as a unified interface, allowing models to query and identify resources without needing to understand the specific APIs or data structures of each platform. This is achieved by connecting to Firefly, which manages the connections to the underlying cloud and SaaS accounts. The AI model can then use natural language queries to find specific resources, such as "Find all 'ubuntu-prod' EC2 instances in AWS account 123456789012." This simplifies the process of gathering contextual information for AI models, enabling them to make more informed decisions and automate tasks across different environments. For example, an AI model could use this feature to identify all running databases before initiating a backup process.

Resource Codification to Infrastructure-as-Code

Firefly MCP can convert discovered cloud resources into Infrastructure-as-Code (IaC) formats like Terraform. This allows AI models to not only identify resources but also understand their configuration and dependencies in a structured, machine-readable format. The AI model can request the IaC representation of a resource, enabling it to analyze, modify, or recreate the resource using standard IaC tools. For example, after discovering an EC2 instance, the AI model can request its Terraform configuration, modify the instance type, and then apply the changes using Terraform. This feature is particularly useful for automating infrastructure management tasks and ensuring consistency across environments. The codification process leverages Firefly's ability to translate resource configurations into various IaC formats.

Secure Authentication via API Keys

Firefly MCP uses API keys (FIREFLY_ACCESS_KEY and FIREFLY_SECRET_KEY) for secure authentication, ensuring that only authorized AI models can access and manipulate cloud resources. This security mechanism protects sensitive infrastructure data and prevents unauthorized access. The API keys are used to authenticate the AI model with the Firefly platform, which then handles the authorization and access control for the underlying cloud and SaaS accounts. This approach simplifies the security management for AI-driven infrastructure automation, as the AI model only needs to manage the API keys, while Firefly handles the complexities of cloud security. The keys can be passed either as environment variables or command-line arguments when starting the Firefly MCP server.