GraphQL and AI with Apollo CTO Matt DeBergalis

Key takeaways

GraphQL provides a powerful and flexible way to interact with APIs, especially in large-scale environments.
Apollo enables developers to adopt GraphQL incrementally by layering it over existing REST APIs and databases.
Strong typing and schema documentation in GraphQL facilitate better API interaction for AI agents.
The hype around GraphQL is evolving, focusing now on its practical applications in enterprise environments.
Community involvement is essential for the development and improvement of GraphQL tools.

Summary

Introduction to GraphQL and Apollo

TJ Vintol and Paige Neatring House welcome Matt DeBergalis, the CTO and co-founder of Apollo, as the special guest of this episode. They dive into the evolution of GraphQL and how Apollo aims to make it more accessible to developers and AI agents. Matt shares the history of Apollo and ties it back to his experience with Meteor, emphasizing the need for a modern way to build applications that communicate efficiently with multiple APIs.

Apollo's Approach to API Integration

Matt discusses how Apollo differs from other GraphQL implementations, particularly focusing on its incremental adoption for teams. With the ability to overlay GraphQL on existing REST APIs and databases without rewriting the backend, Apollo simplifies the transition for large organizations with complex codebases and numerous APIs. This makes GraphQL appealing as it provides a declarative approach to data fetching, minimizing boilerplate and enhancing efficiency.

The Value of GraphQL for AI and MCP

The conversation shifts to the integration of GraphQL with AI, particularly in developing AI agents and the context of API call orchestration. Matt explains how GraphQL’s strong typing and schema documentation enhance the AI's ability to interface with APIs, promoting reliable and predictable interactions which are essential in a regulated industry. He emphasizes the importance of making GraphQL a practical fit for AI, especially for companies in sectors such as finance and healthcare.

Navigating the Hype Cycle of GraphQL

TJ and Matt reflect on the past popularity of GraphQL and its current standing in the tech landscape. There has been a normalization of its usage, with recognition that GraphQL shines particularly when used for larger projects involving multiple APIs, rather than isolated cases. They discuss how Apollo focuses on facilitating GraphQL adoption by managing existing APIs' challenges, thus meeting the needs of engineers who demand effective solutions.

Community Engagement and Future Directions

Matt stresses the importance of community feedback in shaping Apollo’s development. He highlights Apollo's goal to continuously improve the developer experience and mentions upcoming features aimed at easing the integration of GraphQL with existing tech stacks. The episode concludes with an encouraging note that while challenges lie ahead, there is great potential in advancing the AI development stack with GraphQL's capabilities.

Show notes

CTO and co-founder of Apollo, Matt DeBergalis, joins us on this episode to talk about how GraphQL has continued to evolve over time, and how Apollo is focused on making it more accessible for developers and AI agents than ever before.

For those less familiar with Apollo and GraphQL, Matt shares the history of both, including lessons he learned from founding the company Meteor that helped him while building Apollo.

We discuss how even though GraphQL is making less headlines than it was a few years ago, it’s really begun to find its niche within larger organizations that have hundreds or even thousands of APIs and databases underpinning their many applications, and how Apollo has continued to evolve so that it can support APIs, serverless functions, and SQL- or no SQL-databases, with little extra code needed to make these different data sources work together.

Matt also highlights the benefits of a GraphQL schema for AI agents and MCP servers, sharing how the agents are generally very good at parsing the schemas and understanding how to leverage queries against the interface to retrieve the data they need. While we’ve had tech stacks in the past like LAMP and MERN, this new addition of AI to the development mix provides a unique opportunity to redefine the stack once more, and GraphQL could be a very good piece to include.

Special Guest

Matt DeBergalis, CTO and co-founder of Apollo

Relevant Links:

What Makes Us Happy this Week:

Paige - The Pitt TV series
TJ - I Think I Was Murdered
Matt - Teaching my daughter to ride a bike

Thanks as always to our sponsor, the Blue Collar Coder channel on YouTube. You can join us in our Discord channel, explore our website and reach us via email, or talk to us on X, Bluesky, or YouTube.