Top 3 MCPs For AI Coding

Top Three MCP Servers for AI-Assisted Coding

Key Concepts:

• MCP (Multi-Code Provider/Multi-Capability Provider): Servers that provide AI models with access to information or the ability to perform actions outside of the coding environment.
• API Documentation: Technical documentation detailing how to interact with a software service (like Stripe).
• CLI (Command Line Interface): A text-based interface for interacting with a computer system, often used for backend operations.
• Tech Stack: The collection of technologies used to build and run an application.
• Agent: In this context, the AI model assisting with coding.

I. Introduction & The Value of MCPs

The video highlights the significant benefits of utilizing MCP servers when coding with AI tools like Cloud Code, Cursor, or Windsurf. The core argument is that providing AI models with more information and capabilities leads to substantially improved code outputs. The speaker emphasizes that the quality of code generated is directly correlated to the context available to the AI. As stated, “the more information an AI model has when coding, we get better outputs.”

II. Stripe: The Billing Integration MCP

The first recommended MCP is Stripe (or its equivalent, like PayPal, if Stripe isn’t used). Stripe is presented as crucial for ensuring accurate and up-to-date billing implementation.

• Functionality: The Stripe MCP integrates Stripe’s entire API documentation directly into the AI’s coding environment. This allows the AI to access the latest best practices for billing setup and avoid errors caused by outdated information.
• Installation: Installation involves a simple “install” process within Cursor (or similar platforms), followed by connecting and authorizing the Stripe account.
• Importance: The speaker stresses the critical nature of accurate billing, stating, “You’re dealing with people’s money, so you want to make sure your AI model has the most context.” A billing error is considered more damaging than other potential software bugs.
• Example: The speaker illustrates the benefit by mentioning the ability to ask the AI to “create a coupon via the API” or access information about “checkout sessions” directly within the coding environment.

III. Zapier: The Action-Oriented MCP

Zapier is presented as the second essential MCP, focusing on enabling the AI to perform actions outside of the coding environment.

• Functionality: Zapier connects to over 8,000 applications, allowing the AI to automate tasks like creating Jira tickets, sending emails, or posting messages to Slack.
• Installation: Installation involves clicking a link in the video description, selecting the coding platform (e.g., Cursor), and authorizing the connection.
• Example & Case Study: A specific example is provided: automatically creating a Jira ticket for adding an FAQ to a landing page simply by chatting with the AI. The speaker demonstrates this by typing “Want to set up a Jira ticket or FAQ” and the AI automatically creates the ticket.
• Integration with GitHub: The speaker also mentions Zapier’s ability to integrate with GitHub PRs (Pull Requests).
• Distinction from Stripe: Zapier is positioned as complementary to Stripe – Stripe provides knowledge about billing, while Zapier enables actions across various applications.

IV. Tech Stack Specific MCP: GCP as an Example

The third MCP recommendation is tailored to the individual developer’s tech stack (e.g., Google Cloud Platform (GCP), Microsoft Azure, Amazon Web Services (AWS)).

• Functionality: This MCP provides access to API documentation and the ability to directly interact with the backend environment via CLI.
• Discovery: The speaker advises searching for “[Your Tech Stack] MCP [Your Coding Platform]” (e.g., “AWS MCP Cursor”) to find relevant integrations.
• Benefits: This allows the AI to perform tasks like checking for errors in the backend, without requiring manual login to the admin dashboard.
• Example: The speaker provides a personal example of using the GCP MCP to ask the AI to “check all of the errors that incurred in the last 24 hours for all my functions” and receiving the information directly in the chat.
• CLI Importance: The combination of CLI and MCPs is highlighted as particularly powerful, allowing the AI to execute commands and retrieve information directly from the backend.

V. Synthesis & Three Fundamental MCP Types

The video concludes by reiterating the importance of using MCPs to improve code quality and efficiency. The speaker identifies three fundamental types of MCPs:

1. Knowledge-Based MCPs: (e.g., Stripe) – Provide access to API documentation and best practices.
2. Action-Oriented MCPs: (e.g., Zapier) – Enable the AI to perform actions across various applications.
3. Tech Stack MCPs: (e.g., GCP) – Allow direct interaction with the backend environment via CLI.

The speaker emphasizes that while coding without MCPs is possible, the resulting code will likely be inferior. The ultimate goal is to minimize “bloat and bad code” by leveraging the power of these integrations. As the speaker states, “Why the heck are you still writing bad code?”