Reading Time: 8 minutes

Good documentation is one of the most important and overlooked aspects of a well crafted API. If written correctly, it provides the foundation for comprehension for the entirety of the APIs life cycle, with updates that reflect the growth in capability. 

Unfortunately, this is rarely a reality. Documentation is often seen as an afterthought; something hastily cobbled together to tick a box come GA, and neglected as a laborious part of any on-going release process. Inconsistency can creep in at best and at worst, documentation is foregone entirely. 

Thanks to the ongoing widespread adoption of AI tooling, the software development lifecycle is reorienting massively to focus on pace of delivery. Speed is the name of the game when it comes to developing and deploying applications. You’d be forgiven for thinking that documentation may prove a casualty of this shift, falling on the wayside as a sacrifice on the altar of efficiency. 

Rather than fading into obscurity, documentation is going to be exceedingly important. It’ll be the first port of call for both developers looking to understand how to leverage new AI tooling, as well as autonomous AI agents that will scan documentation as the North Star in order to comprehend how to take action in third-party systems.

Thanks to the power of Generative AI, the ability to create consistent documentation at scale has never been easier. That’s why MuleSoft is delighted to announce the general availability of Einstein for Anypoint Exchange: AI Documentation. 

Building on API reusability with Generative API Documentation

Clear, consistent, and complete documentation is the cornerstone of API reusability. Einstein for Anypoint Exchange addresses this directly by providing auto-generated documentation that makes it easier for Integration Developers and Admins to understand and utilize APIs. This leads to increased adoption and more efficient use of assets. 

Our phased rollout ensures that this feature will also improve governance and the documentation for existing assets over time. 

With the option for both agentic creation and manual editing, you maintain control while leveraging the power of Einstein. With Einstein for Anypoint Exchange, MuleSoft customers can now automatically generate documentation based off of the API specifications published to Anypoint Exchange. The benefits include: 

  • Eliminate common documentation pitfalls such as inconsistency, deprioritization and incompleteness. 
  • Automatically generate standardized documentation for REST APIs. 
  • Reduce time-to-value for your business critical IT assets and enhance asset internal marketability, while also improving implementation accuracy. 

Key features of Einstein for Anypoint Exchange

Einstein for Anypoint Exchange is packed with features designed to make API documentation a seamless part of deploying APIs to Anypoint Exchange: 

  • Automatic AI-generated documentation: Users can generate documentation with Einstein via the Exchange UI for REST APIs without documentation 
  • REST support: Initially supports REST API types.
  • Draft view in Exchange: Users can view a draft of the Einstein generated documentation in the asset details page.
  • Manual editing: Users have the ability to manually edit and refine the generated documentation
  • Standardized template: Ensures consistent formatting and includes key information such as title, purpose, description, features, endpoints, and examples.
  • Functional requirements: Includes automatic extraction of information, consistent templates, manual enrichment, AI-based enrichment, and AI-based quality management.

What’s next?

We have an exciting roadmap for Einstein for Anypoint Exchange, including:

  • Other asset types: Documentation generation for existing assets without documentation, and integration with governance conformance status.
  • Autonomous agent documentation management: User interaction with an agent to manage documentation for existing assets.
  • Quality evaluation: Asset quality evaluation with respect to governance.
  • Support for multiple API types: Support for SOAP and Async APIs, and more.

Good documentation ultimately fosters a shared understanding between API providers and consumers, signaling quality and driving adoption. This ethos has been championed by MuleSoft as part of our broader commitment to API Management and it rings true in the world of Agentic AI; API Documentation will be one of the key assets consumed by AI systems in order to understand the art of the possible, and to convey those possibilities to human user building out agentic use cases. 

Documentation parsability is going to be paramount for the future of Agentic Transformation and we’re excited to continue innovating in this area. Stay tuned for more updates to Einstein for Anypoint Exchange. 

Please note: To use Einstein for Anypoint Exchange, users will need Einstein credits, accessible via a connected Salesforce org.