Changelog Glossary
Plain definitions of the terms that come up when teams publish product updates: changelogs, release notes, changelog widgets, llms.txt, and the Model Context Protocol.
This glossary defines the vocabulary of product-update publishing: what a changelog is and how it differs from release notes, what a changelog widget does, what the llms.txt convention asks a site to publish, and how the Model Context Protocol lets AI agents write updates.
Changelog
What is a changelog?
A changelog is a dated, public record of the changes made to a product, listed newest first. Software teams publish one so customers, sales, and support can see what shipped without asking engineering. A changelog entry usually names the change, explains who it affects, and links to documentation.
Read the definition →Release notes
What are release notes, and how do they differ from a changelog?
Release notes are a written summary of what a specific release contains, published when that release ships. A changelog is the running list of those releases over time. Release notes describe one version in depth; a changelog is the archive readers scroll through.
Read the definition →Changelog widget
What is a changelog widget?
A changelog widget is an embeddable component that shows recent product updates inside an application, usually behind a bell icon or a "What's new" link. It pulls from the same entries as the hosted changelog page, so one published update appears in both places.
Read the definition →llms.txt
What is llms.txt?
llms.txt is a plain-text file at the root of a website that gives large language models a curated, Markdown-formatted map of the site's most useful content. It was proposed by Jeremy Howard in 2024 and is a convention, not a standard: no model is obliged to read it.
Read the definition →Model Context Protocol (MCP)
What is MCP?
The Model Context Protocol is an open standard, introduced by Anthropic in November 2024, that lets AI assistants connect to external tools and data sources through a common interface. An MCP server exposes tools; an MCP client such as Claude Code or ChatGPT calls them on the user's behalf.
Read the definition →