AGENTS.md
AGENTS.md is an open convention: a markdown file placed in a code repository that gives AI coding agents project-specific instructions, such as how to build, test, and navigate the codebase and which conventions to follow. It works like a README written for machine collaborators instead of human ones.
Definition
The file sits at the repository root (or in subdirectories for monorepos) and contains the operational knowledge an autonomous coding agent needs: setup commands, test invocations, style rules, directory explanations, and warnings about fragile areas. Tools including OpenAI Codex, Cursor, and many other coding agents read the file automatically before acting on a project.
AGENTS.md matters beyond engineering because it is an early, widely adopted example of content written explicitly for AI consumption. The same principle now spreads to marketing surfaces: structured, unambiguous, machine-readable instructions outperform prose that assumes a human reader who can infer context.
Why it matters for AI visibility
For brands, AGENTS.md is a signal of where content is heading: audiences increasingly include software agents, not just people. Teams that learn to write for machine readers in their repositories develop the same muscle their marketing sites need for AI search, where clear structure and explicit statements decide whether a model can use, and therefore cite, the content.
Related terms
LLMs.txt
LLMs.txt is a proposed convention: a plain markdown file served at the root of a website that gives large language models a curated summary of the site's most important content. It helps AI systems find, understand, and use the right pages at answer time instead of parsing cluttered HTML.
GEOLLMs-full.txt
LLMs-full.txt is a proposed extension to LLMs.txt that includes the full text of content publishers want included in LLM training datasets, allowing direct provision of training data to AI systems.
GEOAI Web Crawlers
AI Web Crawlers are automated systems deployed by LLM providers and AI companies to index and retrieve web content for LLM training, inference retrieval, or answer synthesis processes.
AIFunction Calling (Tool Use)
Function Calling is the capability of LLMs to call external functions or APIs as part of generating responses. It enables agents to take actions, query databases, or access real-time information beyond the model's training.