Skip to content

News alerts in C# with Google News RSS search feeds and Discord

The goal: a feed of recent news on a handful of market-related keywords, pushed to a Discord channel, so there’s no need to check a news site every day. Google News has an RSS search endpoint that does most of the work.

Put your search terms in the q parameter:

https://news.google.com/rss/search?q=technology

The response is an RSS 2.0 feed. Each <item> has a title (headline followed by ” - Publisher”), a link to a Google News redirect URL, a guid, and a pubDate in RFC 822 format, such as Fri, 13 Nov 2026 22:57:09 GMT. You can subscribe to that URL in any feed reader, or fetch it from code as below.

Google doesn’t document this endpoint as a public API, so the format could change without notice. Check Google’s terms before republishing its content anywhere.

For each search term, the app:

  1. Fetches the RSS feed.
  2. Looks at the first five items and skips anything older than two days.
  3. Checks Elasticsearch for a document with the item’s guid. If one exists, the story has already been sent.
  4. Otherwise, posts the title, date and link to a Discord webhook, then stores the guid in Elasticsearch.

Elasticsearch is overkill for “have I seen this ID before?”, and SQLite or a text file would do the same job. I used it because I already had a hosted cluster. Using the guid as the document ID makes the check an exact lookup rather than a search.

The full original source is at https://github.com/FriendlyUser/news-alert/blob/master/News.cs. The repo also has a checkPrices stub for stock price alerts that never got past printing the ticker, so it’s left out here.

This is a .NET 8+ console app (dotnet new console) with no extra packages. Set DISCORD_WEBHOOK to a Discord webhook URL and ES_INSTANCE to the Elasticsearch base URL.

using System.Net.Http.Json;
using System.Xml;
var http = new HttpClient();
string? webhook = Environment.GetEnvironmentVariable("DISCORD_WEBHOOK");
string? esInstance = Environment.GetEnvironmentVariable("ES_INSTANCE");
if (webhook is null || esInstance is null)
{
Console.WriteLine("Set DISCORD_WEBHOOK and ES_INSTANCE first.");
return;
}
string[] searchItems = ["hydrogen", "analyticGPT", "natural gas", "recession", "Tech layoffs", "Bard AI"];
foreach (string searchText in searchItems)
{
string url = $"https://news.google.com/rss/search?q={Uri.EscapeDataString(searchText)}";
string xml;
try
{
xml = await http.GetStringAsync(url);
}
catch (HttpRequestException e)
{
Console.WriteLine($"{searchText}: {e.Message}");
continue;
}
var doc = new XmlDocument();
doc.LoadXml(xml);
XmlNodeList items = doc.GetElementsByTagName("item");
for (int i = 0; i < Math.Min(5, items.Count); i++)
{
XmlNode item = items[i]!;
string? title = item["title"]?.InnerText;
string? link = item["link"]?.InnerText;
string? guid = item["guid"]?.InnerText;
string? pubDateText = item["pubDate"]?.InnerText;
if (title is null || link is null || guid is null || pubDateText is null)
{
continue;
}
// DateTimeOffset keeps the GMT offset; DateTime.Parse would convert to local time
DateTimeOffset pubDate = DateTimeOffset.Parse(pubDateText);
if (DateTimeOffset.UtcNow - pubDate > TimeSpan.FromDays(2))
{
continue;
}
// GET returns 404 if the document (or the index) doesn't exist yet
string docUrl = $"{esInstance.TrimEnd('/')}/post/_doc/{Uri.EscapeDataString(guid)}";
using var existing = await http.GetAsync(docUrl);
if (existing.IsSuccessStatusCode)
{
Console.WriteLine($"Already sent: {title}");
continue;
}
string message = $"{title}\n{pubDateText}\n{link}";
using var sent = await http.PostAsJsonAsync(webhook, new { content = message });
sent.EnsureSuccessStatusCode();
// PUT creates the "post" index on first use if auto-create is enabled
using var indexed = await http.PutAsJsonAsync(docUrl, new { guid, title, link });
indexed.EnsureSuccessStatusCode();
}
}

If your cluster needs authentication, set http.DefaultRequestHeaders.Authorization (basic auth or an API key) rather than relying on credentials embedded in ES_INSTANCE.

  • The old version read items[i] for i < 5 without checking how many items came back. Feeds for niche terms can return fewer than five, so the loop is now capped at items.Count.
  • The old version parsed pubDate with Convert.ToDateTime, which returns local time, and compared it with DateTime.UtcNow. On a machine outside UTC the two-day window was off by the timezone offset.
  • Discord webhooks are rate limited. A run that finds dozens of new stories at once can get 429 responses, so add a delay between posts if you widen the search list.
  • Google News links are redirects, not the publisher’s URL. If you need the real article URL, follow the redirect or use the <source> element for the publisher’s site.

© 2026 David Li · Home · Blog · RSS ·Classic site · Built with Astro, Starlight and three.js · Logos:Font Awesome Free (CC BY 4.0)