> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getsurfaced.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Articles

> Your content library: reviewing drafts, the fact-check, revisions, and archiving.

Articles is everything that has been written, at whatever stage. Each one is built from blocks (paragraphs, headings and the rest) and keeps its own revision history.

## Two different statuses

Every row carries two, and they are independent:

**Draft checks.** Whether the quality checks have run and what they found: not checked, queued, checking, checks passed, changes needed, or could not finish.

**Publication.** Not published, published, live URL linked, or **draft differs from live**. The last one matters: the blog serves a snapshot taken at publish time, so editing a published article changes the draft and leaves the live page alone until you update it.

A **review needed** chip appears when the pipeline shipped something on a judgement call that nobody has looked at since.

## Reviewing a draft

A draft arriving from a [brief](https://docs.getsurfaced.ai/content/briefs.md) run is finished text, not an outline. Read it for two things: whether every claim about your product holds, and whether it sounds like you.

**Draft checks** gives a score out of 100 and a list of items. Two kinds: a **finding** is serious enough to block publication, a **suggestion** is not. Each can be fixed in place, with help from **Fix with AI**, and then rechecked.

The checks cannot catch a statement that is plausible, unsourced and wrong about your own product. Read for those yourself.

Voice is the last stretch. Brand voice settings get most of the way there; the rest is editing.

## Editing

Block content is edited in the app. Metadata (title, description, excerpt, categories, tags, language) can also be changed through the [API](https://docs.getsurfaced.ai/api/public-api.md).

Every saved change creates a revision you can view and restore. Restoring is version-checked, so it cannot silently overwrite an edit someone else made while you were looking at an older version.

## The publish gate

An article that has not passed its draft checks cannot be published. This is enforced in the database rather than in the page, so it applies however publication is attempted: the app, the API, any integration. There is no override and no "publish anyway", for any role.

The message you get is: *check this page revision and resolve its findings before publishing.*

> [!WARNING]
>
> The gate is evaluated per revision, continuously, not once at publish time. Rechecking a live article and getting "changes needed" takes it off the blog until it passes again, and the URL 404s in the meantime. Recheck a published page when you are ready to fix what comes back, not to see what happens.

**Publish** puts the article on your [hosted blog](https://docs.getsurfaced.ai/publish/hosted-blog.md) and returns the live URL. **Update on blog** pushes later edits to the live page. **Unpublish** removes the post and keeps the article.

## Archiving and deleting

They are not the same thing.

**Archive** is reversible. The article leaves the default list, keeps its history, and the toast offers an immediate undo. Archiving a published article also takes it off the blog, and restoring it does **not** put it back, so a restored article has to be republished.

**Delete permanently** is a hard delete with no undo.

Prefer archiving. It is reversible, and it keeps the record of what you once published.

## Did it work?

Publishing is not the end of the loop. Whether a page is being cited is on [Content performance](https://docs.getsurfaced.ai/insights/content-performance.md), and it takes time: the engines have to re-read your site, and then a tracking run has to ask the questions again.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.