Google News → Markdown

Google News and Google Alerts,
as full-text Markdown

Paste your Google News or Alerts feed. Get every new article as clean Markdown, delivered by webhook. Real URLs, not encrypted redirects. No duplicates, no sidebar junk.

Google News feedsGoogle Alerts RSSFull article textClean MarkdownWebhook deliveryCross-source dedupn8n node

Pay per feed, not per page. Plans from $19/mo.

Before and after

The feed gives you a redirect. We give you the article.

Same story, two very different payloads. One of them your LLM can actually read.

What Google News RSS gives you

<item>
  <title>OpenAI announces new reasoning model - TechCrunch</title>
  <link>https://news.google.com/rss/articles/CBMiW0FVX3lxTE...</link>
  <description>&lt;a href="..."&gt;OpenAI announces...&lt;/a&gt;
    &amp;nbsp;&amp;nbsp;TechCrunch</description>
  <pubDate>Fri, 19 Sep 2026 14:02:00 GMT</pubDate>
</item>

What Markdownfeed sends to your webhook

{
  "title": "OpenAI announces new reasoning model",
  "url": "https://techcrunch.com/2026/09/19/openai-...",
  "source": "TechCrunch",
  "author": "Kyle Wiggers",
  "published_at": "2026-09-19T14:02:00Z",
  "markdown": "# OpenAI announces new reasoning model\n\n
    OpenAI said on Friday that its new model..."
}

The problem

Why your Google News pipeline keeps breaking

Every team that automates news curation hits the same four walls, usually in this order.

Encrypted links

The link is not the article

Google News hands you CBMiW0FVX3lxTE..., not a publisher URL. Your scraper follows it and gets a redirect shell, or nothing at all.

Two-line snippets

There is no body text

The feed carries a headline and a teaser. Your summariser, your classifier and your digest all need the full article.

Duplicates

One story, twenty times

Twenty outlets run the same wire story. Without cross-source dedup your newsletter lists all twenty and your token bill pays for each.

Junk markup

Navigation in your prompt

Cookie banners, related-article rails, newsletter signup boxes and footers all end up in naive Markdown output.

One tool instead of three

Replaces rss.app + Firecrawl + your dedup script

Same output, one bill, no glue code to keep alive when a source changes its markup.

rss.app

$8–100/mo

Firecrawl

$16–83/mo

Your dedup script

your weekends

→

Markdownfeed

from $19/mo

One feed URL. One webhook. Nothing to maintain between them.

How it works

Three steps, then it runs by itself

Set a feed up once. New articles keep arriving until you turn it off.

Paste your sources

A Google News search URL, a Google Alerts RSS link, or both. Up to 10 sources on Starter, 200 on Agency.

news.google.com/rss/search?q=...

We watch them for new items

Every run we resolve the real article URL behind the redirect, fetch the page in a real browser and strip everything that is not the article.

interval=1h · new_only=true

You get the article

A webhook POST per new article, or pull from the JSON and RSS endpoints. Title, author, source and publish date come with it.

POST /your-endpoint

API

One POST to create a feed. Articles arrive on their own.

A REST API and a webhook. No SDK to learn, no per-site parsers to write, no cron job to babysit.

curl -X POST https://api.feedmarkdown.com/v1/feeds \
  -H "Authorization: Bearer $MARKDOWNFEED_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "AI news",
    "sources": [
      "https://news.google.com/rss/search?q=openai",
      "https://www.google.com/alerts/feeds/1234/5678"
    ],
    "interval": "1h",
    "webhook_url": "https://hooks.yourapp.com/articles"
  }'

Need bulk feeds or a private deployment?

Same API, higher limits, per-source SLA.

No-code

Drops straight into n8n, Make and Zapier

Install the Markdownfeed node, add the New article trigger, and every story arrives already parsed. No HTTP Request node, no HTML parsing, no Remove Duplicates chain.

New article trigger fires once per story
Fields mapped: title, source, author, published_at, markdown
Works the same over a plain webhook in Make or Zapier
Ready-made templates for digests, Slack and LinkedIn

Workflow · AI newsletter

Markdownfeed

New article

OpenAI

Summarise markdown

Filter

Score > 7

beehiiv

Append to draft

Same story, ten sources

14:02Reuters
webhook sent
14:09CNBC
deduplicated
14:11Yahoo Finance
deduplicated
14:24Benzinga
deduplicated
15:40The Verge
webhook sent

Deduplication

One story, one webhook call

We dedupe on URL and on content, across every source in the feed. When a wire story runs in eleven places, your endpoint hears about it once.

Canonical URL matching after redirect resolution
Content fingerprinting for syndicated and rewritten copy
Re-runs never resend an article you already received
Per-feed history you can replay for the last 30 days

Use cases

What teams build with Markdownfeed

Same engine underneath. What changes is the source list and where the webhook points.

NEWSLETTERS

Daily and weekly AI digests

Pull 100 stories a day into a digest without a four-hour reading session. Feed the Markdown straight into your summarisation prompt.

Article body arrives ready for the model
Duplicates removed before they cost tokens
Publish date and source for attribution lines

AGENCIES

One feed per client

Run news automation for a roster of clients from one account. Separate feeds, separate webhooks, one invoice.

Up to 200 sources on the Agency plan
Per-source health so a dead site is visible
White-label the output, we stay invisible

AI PRODUCTS

Ingestion for news apps

Skip building the scraping layer. Point a feed at your source list and get clean documents into your index as they publish.

JSON, RSS or webhook delivery
Stable schema, no per-site parsers
Backfill the last 30 days on feed creation

MONITORING

Alerts that carry the story

Google Alerts tells you something happened. Markdownfeed tells you what it said, in the same webhook call.

Google Alerts RSS in, full text out
Route to Slack, Notion or your own store
15-minute checks on the Agency plan

Pricing

Priced per feed, not per page

Scraping 3,000 articles a month costs the same as scraping 300. No credits to top up, no surprise overage.

Starter

$19/mo

One newsletter, one topic.

10 sources per feed
Checked every 6 hours
Webhook, JSON and RSS delivery
Cross-source deduplication
7-day article history

Questions we get a lot

Does this work with sources that are not in Google News?+

Google News and Google Alerts are supported today, and they cover most publisher content. Direct site sources and plain RSS are on the roadmap — tell us which sites you need and we will prioritise them.

How is this different from just scraping the RSS feed myself?+

Google News links are encoded redirects, so the first problem is finding the real article URL. After that you need a renderer that survives JS-heavy news sites, a content extractor that drops navigation and related-article rails, and dedup across sources. That is the product.

What about paywalled articles?+

We return whatever the page serves to an anonymous visitor. When a body is truncated behind a paywall we flag it in the payload so you can skip it instead of summarising three paragraphs.

How quickly do articles arrive?+

Between 15 minutes and 6 hours after publication, depending on your plan's check interval. Delivery to your webhook happens within seconds of us fetching the article.

Is scraping news articles legal?+

We fetch publicly available pages the same way a browser does and hand the text to you. What you publish from it is your call — most customers summarise and link rather than republish in full.

Can I move an existing rss.app and Firecrawl pipeline over?+

Yes, and that is the common case. Send us your current source list and we will set up the equivalent feed so you can run both in parallel before switching the webhook.

Point a feed at Google News.
Get articles, not redirects.

Set up your first feed in a few minutes, or send us your source list and we will build it for you.

Google News and Google Alerts, as full-text Markdown