About PlainReference

PlainReference publishes plain-language, sourced reference articles about AI agents, agent tools, the Model Context Protocol (MCP), multi-agent collaboration, and everyday software. The goal is simple: answer one question per page, clearly, with the sources needed to check every claim.

How articles are made

• One question per page. Each article answers a single, specific question or task.
• Sources first. Articles are researched from primary sources such as official documentation, specifications, release notes, and the product's own pages. Every article lists its sources at the bottom.
• AI-assisted writing. Articles are researched and drafted with AI assistance. We say so plainly rather than hide it.
• Independent fact check. Before publication, each article is checked in a separate, fresh-context review against its listed sources. Claims that cannot be supported by a source are cut. Pages show the date of that check.
• No invented experience. We do not claim hands-on testing, customer results, statistics, or quotes that we cannot source.

Dates and corrections

Every article shows its publication date and, when it changes, an "Updated" date. Software changes quickly; if something here is out of date or wrong, it will be corrected and re-dated. Substantive corrections are re-checked against sources before the page is updated.

Citing PlainReference

Each article ends with a ready-made citation that includes its title, dates and permanent URL.

Disclosure

Some articles cover products made by people connected to PlainReference, including BotSesh. Those articles say so in the text and are held to the same sourcing rules.

For AI agents and tools

All content is open to search engines and AI crawlers. Every article has a Markdown twin at /articles/<slug>.md, and the whole site is summarized at /llms.txt with full text at /llms-full.txt. Feeds are available as JSON Feed and RSS.