# Introducing SaturnDocs

A site built only for people leaves these agents to guess at the answers they need.

×

Mission log · By SaturnDocs

A site built only for people leaves these agents to guess at the answers they need.

Published July 10, 2026 Updated September 6, 2026

 ![](https://saturndocs.com/assets/space/saturn-vista-liftoff-960.webp)

People still read documentation. They read it to evaluate a product before they buy and to use it after. Coding agents read it too. Cursor, Claude Code, and ChatGPT read the pages before and after a purchase decision. So do the agents a company builds itself. A site built only for people leaves these agents to guess at the answers they need.

When an agent cannot find a clean answer, it does one of two things. It guesses from whatever it can scrape, or it tells the person it is helping that the answer is unclear. Either outcome reflects on the product behind the docs.

A site has to do different work for each kind of reader. For people it needs search, guides, an interactive API reference, and an assistant that answers questions from the pages and cites them. For agents it needs surfaces they can fetch and parse. Those surfaces include llms.txt, llms-full.txt, a Markdown version of every page, an MCP server, WebMCP, and a skills index. A site that serves people well but gives agents nothing leaves the agents to scrape rendered HTML or skip the docs entirely.

Reader

What the site has to give them

People

Search, guides, an interactive API reference, an assistant that answers from the pages and cites them

Agents

llms.txt, llms-full.txt, a Markdown version of every page, an MCP server, WebMCP, a skills index

## What every site includes

Every SaturnDocs site has both sets of surfaces. It has search, guides, an interactive API reference, and an assistant that cites its sources, for the people who read in a browser. It has llms.txt, llms-full.txt, a Markdown version of every page, an MCP server, WebMCP, and a skills index, for the agents that fetch and parse. Hosting is part of the site, and you can set pages to be public, restricted to your customers, or private.

The agent-facing surfaces and the person-facing pages come from the same source. When a page changes, its Markdown version changes with it. The skills index and the MCP server point at the same content the browser shows.

The agents do the writing. They build the first version of the site from what already exists. They read your repository, your API spec, your old docs, your notes, or they take a call with your team. Then they publish the site on your domain. They update the docs after every product release. They fix what you flag. The work that a documentation team would do by hand, our agents do from the same sources, and they repeat it after each release.

You do the parts that require your judgment. You send the product context. You answer questions where the sources disagree, because the agents cannot decide which source is right. When the agents find a gap, they ask a question and wait for your answer rather than filling it with a guess. You approve each change before it publishes.

## What stays with you

The files are plain files. They are readable, editable, and exportable. If you leave SaturnDocs, you keep them. Because the files are plain Markdown, you can move them to another system without a migration project, and you can edit any sentence by hand. Your product knowledge stays yours because you supplied it. Your approval stays yours because nothing publishes without it.

The agents can only work from what exists and from what you tell them. Where two sources conflict, they stop and ask. That is a limit on what they can do alone. It is also the point at which your judgment decides.

We offer a limited-time Early Adopter plan at $100 per month, with the price locked for one year and migration included. Standard serves one product team, and Enterprise supports larger documentation programs. Every plan includes the complete platform. Contact us about Standard or Enterprise.

## What SaturnDocs does

SaturnDocs is a documentation platform built for the AI agents that now read your docs, and for the people who still do. Our agents write the first version of the site from your repository, your API spec, your old docs, your notes, or a call. They publish on your domain, update the docs after every release, and fix what you flag. You supply the product context and approve each change before it publishes. Send us your repository and your API spec, and we will show you the first version of your site before you commit to anything.

See more SaturnDocs in your Google Search and AI results.

[Make SaturnDocs a preferred source on Google](https://www.google.com/preferences/source?q=saturndocs.com)

## Related articles

August 10, 2026

### [The work a docs subscription leaves on your team](https://saturndocs.com/blog/results-as-a-service/)

The subscription does not buy the work that fills those pages.

July 28, 2026

### [Documentation from the material you already have](https://saturndocs.com/blog/what-done-for-you-means/)

You already hold the raw material for a documentation site.

Your docs, handled

## Ready to take documentation off your list?

Hand over the docs you have. The agents publish, update, and keep every page current.

[Send us your docs](https://saturndocs.com/contact/?intent=docs)

Canonical URL: https://saturndocs.com/blog/introducing-saturndocs/
