GraphQL as an ‘essential protocol’ for AI-API orchestration

A new product will establish the graph-based industry standard for secure, orchestrated access to APIs in the age of AI… so says Apollo GraphQL, a graph-based API orchestration company.

The organisation has now announced the Apollo MCP Server, a technology service that connects graph APIs to AI models using the Model Context Protocol (MCP). 

MCP has emerged as (arguably) a “go-to standard” for giving AI models access to data and capabilities beyond their training. With this new addition to Apollo’s toolkit, developers can now harness GraphQL’s power when building AI applications that need seamless AI access to data and services.

We know that firms are working to provide AI applications with appropriate access to their APIs. 

GraphQL, the emerging standard for declarative API orchestration, offers AI application builders a seamless path to quickly and safely build API-powered agentic software. 

Getting data from APIs into the right shape for AI consumption involves complex challenges around discovering, sequencing, security gating, retrying, parallelising, caching and numerous other functional and non-functional requirements. 

“With the Apollo MCP Server, we’re establishing GraphQL as the essential protocol for AI-API orchestration,” said Matt DeBergalis, CTO and co-founder of Apollo GraphQL. “While MCP is still evolving and perspectives on implementation will continue to mature as the technology develops, what’s clear today is that GraphQL addresses the needs of AI by allowing teams to focus on building new AI-powered business capabilities, not rehashing endpoints specifically for AI integration. Safe, secure and performant AI needs graph-based API orchestration.”

GraphQL’s declarative, schema-driven approach to API orchestration provides these capabilities out-of-the-box, making it suited to serve as the foundation for MCP implementations, especially in an AI landscape where speed, security and agility are paramount.

By implementing the Model Context Protocol, the Apollo MCP Server enables teams to move at AI speed while maintaining appropriate governance. It also simplifies deployment, with options to start in a local environment via Apollo’s Rover CLI or deploy as a containerised service for seamless integration with your cloud infrastructure.

Apollo MCP Server lets developers turn GraphQL schema or operations into MCP tools without writing a single line of procedural code. It works with any combination of unmodified REST APIs via Apollo Connectors as well as native GraphQL services.