These 10 posts are the latest from the Zylosystems' blog.
AI Agents are evolving beyond simple language models into autonomous software systems that can interpret goals and execute actions on their own. This shift is opening up new opportunities in the field of technical documentation as well. If ReadMe and Mintlify represent the first generation of automation in documentation operations, Zylo AI introduces the next generation of documentation agents — systems that understand both source code and customer inquiry data to create truly “living documents.”
Most engineering-driven teams struggle with documentation that doesn’t keep pace with product growth. In this case study, we walk through how Zylosystems helped Notifly transform their static MDX docs into a fully structured, OpenAPI-powered, developer-friendly experience using Mintlify. From rewriting content for Korean developers to building a single unified API spec, this project became the foundation for truly readable documentation—and marked our very first customer success.
At my San Francisco demo of Zylo-Docs, engineers asked why they’d need this when Swagger exists. But technical writers were intrigued by how AI could actually help fill documentation gaps. That contrast opened my eyes to the real opportunity: building AI-powered tools tailored for technical writers.
On August 19 at Elasticsearch’s San Francisco headquarters, we joined the MCP Hack Night to explore how mcp-use and Elasticsearch can power smarter server discovery. By storing MCP server metadata in Elasticsearch and querying it with the Search SDK, we experimented with dynamically connecting only the tools an MCP Agent needs—an approach that could reshape how documentation leverages context and consistency.
From the very first version of zylo-docs, Zylosystems came to Silicon Valley to start customer discovery and validate the product in the US market. During our time in San Francisco, we shared demos, met with founders and technical writers, and built connections to shape the future of developer documentation.
OpenAPI 3.2 is nearing release, with over 94% of issues resolved. This version introduces structured tags with parent relationships, official support for extended HTTP methods like QUERY, and full multipart media type support. Here’s what it means for the future of automated API documentation—and how Zylo-docs is getting ready.
Great API docs are more than just a reference — they’re a gateway to faster adoption and better developer experience. In this post, we explore what makes Fern’s docs so effective and how Zylo-docs goes a step further by integrating directly into your codebase, offering real-time, AI-powered documentation for FastAPI and beyond.
On June 25, Google introduced Gemini CLI, an open-source AI terminal agent that brings Gemini 2.5 Pro to your command line—for free. In this post, we explore what makes Gemini CLI powerful, how it fits into the evolving AI developer ecosystem, and why Zylo Systems is paying close attention to tools that turn AI into the primary user of API documentation.
As LLMs begin to consume and act on documentation like never before, the way we write API docs is fundamentally changing. No longer just for humans, modern documentation must now serve its fastest-growing audience: machines.
An overview of how API guidelines like Google’s AIP (API Improvement Proposals) support scalable API design. Covers principles such as flexibility, continuous evolution, and the importance of tooling—plus practical tips for small teams starting their own API governance journey, based on insights from Google’s internal practices.
