AI coding agents are now part of the developer workflow. Whether we like that shift or hate it, users ask Codex, Cursor, Claude Code, GitHub Copilot, and other tools to install packages, wire examples, migrate code, and explain APIs. For JavaScript data grids, that often means asking an agent to build columns, editors, filters, Pivot views, or Gantt timelines.
We ran into the obvious problem: the docs were written for people. A person can use the sidebar, search, breadcrumbs, examples, and product context. An agent often starts with a URL and a vague instruction. If it cannot find the right entry point quickly, it guesses.
There are several ways to make a site easier for agents to use: an MCP server for live retrieval, skill bundles for structured on-demand context, and llms.txt as a public discovery layer.
That is where llms.txt helps. It is a public, text-first entry point that tells AI agents where the important documentation lives: installation guides, API reference, examples, migration notes, troubleshooting, full text exports, and any richer machine-readable bundles.
The goal is not to “optimize for bots” at the expense of people. The goal is to make sure that when a user asks an agent to build with RevoGrid, the agent can find the same source material a careful developer would read first.
