---
title: "RSS feeds"
description: "Forums, communities and blogs no platform search reaches, read through their RSS or Atom feed."
canonical: https://docs.mentio.dev/platforms/rss
markdown: https://docs.mentio.dev/platforms/rss.mdx
---

# RSS feeds

Forums, communities and blogs no platform search reaches, read through their RSS or Atom feed.

A keyword can read RSS and Atom feeds: a forum, a community, a company blog, a newsletter, a GitHub repository's releases. Paste the feed, or just the site's address: Mentio finds the feed the page advertises (or a usual address such as `/feed` or `/rss`) when the keyword is saved. A post of the feed is a mention of the keyword when it holds the term, with the keyword's [matching rules](/guides/keywords#narrow-a-broad-keyword).

| `platform`                 | `rss`                                                     |
| -------------------------- | --------------------------------------------------------- |
| What is found              | The posts of the feeds a keyword names that hold its term |
| Freshness                  | Every hour                                                |
| Last 30 days on a new feed | Yes: the newest 10 posts holding the term                 |

## Connect a feed

```bash
curl -X PATCH https://api.mentio.dev/v1/keywords/kw_123 \
  -H "Authorization: Bearer $MENTIO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "feeds": [{ "url": "https://forum.example.com" }, { "url": "https://example.com/blog/feed.xml" }] }'
```

`feeds` replaces the whole list, at most 20 a keyword; `[]` disconnects them all (their mentions stay). Each entry is checked when it is saved: an address with no feed behind it, or a site that blocks automated readers, is a 400 that says which. The keyword's `feeds` array shows each feed's own address and title. A keyword may read feeds alone: `platforms: []` searches the term nowhere and keeps its feeds.

A feed belongs to the keywords that named it. Another workspace reading the same feed is polled once with you, but a post of it is never a mention of a keyword that did not name the feed.

## Finding the feed

* **A Discourse forum** (its front page): its latest posts, replies included (`/posts.rss`), rather than its new topics alone. A category page: that category's feed.
* **WordPress, Ghost, Substack, most blogs**: the feed the page advertises, usually `/feed`.
* **Medium**: a tag, a writer or a publication goes straight to its Medium feed.
* **GitHub**: `https://github.com/<owner>/<repo>/releases.atom` for releases.
* **Reddit and X** are refused: Mentio [searches them already](/platforms/reddit), so add the platform to the keyword instead.

## What a mention carries

* `post.title`: the post's title.
* `post.text`: the title, then the post's text as the feed carries it (some feeds carry a summary only).
* `post.url`: the post itself.
* `author`: the author as the feed names them, when it does.

The rest is the same on every platform: relevance, sentiment and intent from the classifier, your triage, and the keyword it matched. See the [mention shape](/api).

## Price

The same as every platform: each matched mention $0.008. Reading a feed costs nothing more.

## Limits

* A feed lists only its newest posts (often 20 to 50). A site that publishes more than that in an hour loses the overflow.
* A post with no date is skipped: nothing would say whether it was published before the keyword started listening.
* Sites behind a bot wall (a Cloudflare challenge, a login) cannot be read.
* The look-back reads what the feed still lists: a feed of 20 posts, none of them older than a day, brings at most the matches among those 20.