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

# Include and exclude selectors

> Precise control over what Reader keeps and drops using CSS selectors.

CSS selectors give you scalpel-level control over what ends up in your markdown. You'll reach for these when `onlyMainContent` isn't precise enough, typically for sites with unusual layouts or very specific extraction requirements.

## includeTags: keep only these

Pass a list of CSS selectors, and Reader keeps only content matching them. Everything else is dropped.

```ts theme={null}
await reader.read({
  url: "https://example.com/product/42",
  includeTags: [".product-title", ".product-description", ".product-specs"],
});
```

Useful when a page has a predictable layout and you want only the relevant fragments: typical for product detail pages, recipe cards, event listings.

## excludeTags: drop these

The inverse. Keep everything, but drop what matches these selectors.

```ts theme={null}
await reader.read({
  url: "https://example.com/blog/post",
  excludeTags: [".newsletter-signup", "#related-posts", "figure.ad"],
});
```

Useful when `onlyMainContent` leaves in something you don't want, or when you've turned it off for a reason but there's still boilerplate to trim.

## Combining them

Both can be passed at once. Reader runs `includeTags` first (narrows to matching content), then `excludeTags` on what remains.

```ts theme={null}
await reader.read({
  url: "https://docs.example.com/api/reference",
  includeTags: ["main"],
  excludeTags: [".feedback-widget", ".edit-on-github"],
});
```

## Selectors that work

Reader supports the full CSS selector grammar that your browser does:

* Element selectors: `article`, `main`, `p`
* Classes: `.content`, `.post-body`
* IDs: `#main-content`
* Attributes: `[role="main"]`, `[data-testid="article"]`
* Descendants: `article .content p`
* Combinators: `main > section`
* Pseudo-classes: `:first-child`, `:not(.sidebar)`

## Tips

* **Inspect the target site first.** Open the URL, use devtools, find the exact selector you want, and copy it.
* **Prefer stable attributes.** `[data-testid="..."]` and semantic tags (`<article>`, `<main>`) are more stable than class names that change with every site redesign.
* **Start broad, narrow down.** Begin with a broad selector, verify Reader returns what you expect, then tighten.
* **Selectors are per-request.** There's no "global" selector list; pass them every time, or wrap your calls in a helper.

## Next

* [Main content extraction](/home/guides/getting-started/main-content): the default that selectors refine
* [Dynamic content](/home/guides/advanced/dynamic-content): when selectors need `waitForSelector`