Forwarding all new articles from RSS to WhatsApp seems like a task suited for automation. However, without clear filtering, the workflow can fill the group with notifications, send irrelevant news, or forward headlines that are already outdated.
RSS is a format that allows a site to provide a structured list of the latest content. Meanwhile, n8n can act as a bridge between RSS, AI services, databases, email, Slack, or WhatsApp. The challenge is not just to move data from one place to another, but to determine when an article is worthy of being forwarded and in what form the information should be received.
The main issue is not data retrieval, but data selection
The RSS Feed Read node in n8n can fetch new articles from a feed. Typically, the received data includes the title, link, publication date, brief description, and sometimes the article's content.
If each item is sent directly to WhatsApp, several issues can easily arise:
- Articles that are off-topic get sent.
- A single source publishes too many updates in a day.
- Sensational headlines may not contain important information.
- Summaries from RSS are too short to be understood.
- Recipients receive messages at inappropriate times.
Therefore, a good workflow should have several layers of decision-making. RSS is just the entry point. The final decision must still pass through filters, grouping, and if necessary, human approval.
Design the workflow before opening n8n
Before creating nodes, first define simple rules. For example, the goal of the workflow is to send news summaries about automation, APIs, and n8n to a work group every afternoon.
Rules that can be used:
- Check the feed every 30 minutes or one hour.
- Only process articles from selected sources.
- Fetch articles containing specific keywords.
- Limit the number of articles sent in one batch.
- Combine several articles into one daily summary.
- Send to WhatsApp only after passing checks or approvals.
Such rules make automation easier to maintain. If all logic is directly input into the AI prompt or into many branching nodes, the workflow will be hard to understand when needs change.
A practical n8n workflow structure
For the RSS to WhatsApp scenario, the basic structure can be created as follows:
- Schedule Trigger runs the workflow at specific intervals.
- RSS Feed Read fetches articles from one or more feeds.
- Filter filters sources, keywords, dates, or languages.
- Code or Set normalizes the format of titles, dates, and links.
- AI or regular rule processing creates summaries and determines relevance.
- Approval allows a human to review the results.
- WhatsApp API or gateway sends the approved messages.
At the initial stage, do not use AI directly. Start with rule-based filters. For example, articles are only processed if their title or description contains words like n8n, webhook, API, or automation.
const keywords = ['n8n', 'webhook', 'api', 'automation'];
const text = `${$json.title} ${$json.contentSnippet}`.toLowerCase();
return keywords.some(keyword => text.includes(keyword));Simple filters are more predictable and do not incur additional costs. AI can be added after the basic flow is running stably.
Use AI for summarizing, not for making all decisions
AI is useful when the RSS description is too short or when the relevance of an article cannot be determined solely from keywords. However, AI should not be given full control to send messages without checks.
A safer instruction is to ask AI to generate a clear structure, for example:
- Article title.
- Summary of up to three sentences.
- Reason the article is considered relevant.
- Importance level: low, medium, or high.
- Source link.
Example instruction:
Summarize the following article in neutral Indonesian. Do not add facts that are not in the text. Indicate whether the article is relevant for readers interested in n8n, API integration, webhooks, or work automation. If the information is insufficient, state that the context is limited.
The separation between summary and important decisions is crucial. AI can help provide recommendations, but business rules can still be defined in n8n. For example, articles with a low importance level only go to the archive, while high-level articles go to the approval path.
It's better to send a digest than individual notifications
Instant notifications are suitable for important alerts, but not ideal for every new article. For content from RSS, the digest format is usually more convenient.
The workflow can collect articles over several hours, then send one message like:
Today's Automation Summary
- First article — brief summary and link.
- Second article — brief summary and link.
- Third article — brief summary and link.
This way, recipients are not disturbed every time the feed changes. n8n can temporarily store results in a database, spreadsheet, or other storage systems before creating a combined message.
Add an approval path for important messages
If messages are to be sent to customers, large groups, or official company channels, add a human-in-the-loop. This means that a human still checks the results before the workflow forwards the message.
The approval path can be created via email, Slack, Telegram, or a simple dashboard. The format can include a summary and two options: approve or ignore.
This separation is important because small errors in the summary can change the meaning of the article. Additionally, some feeds may contain advertisements, opinions, or content that is not yet suitable for public sharing.
Pay attention to WhatsApp rules and limitations
Automatically sending messages to WhatsApp is not just about finding the API endpoint. You need to ensure that the service used is indeed official or meets operational needs, especially if messages are sent to customers.
Check the following:
- Have recipients given permission to receive messages?
- Is there a frequency limit for sending?
- How to handle failed messages?
- Do links and message formats display well on different devices?
- Is article data or recipient numbers stored securely?
For internal use, you may choose a simpler gateway or integration. However, for business communication, aspects of approval, privacy, and platform policies must be part of the design from the start.
What can be done now
Start with a small workflow: one RSS feed, one keyword filter, one storage place, and one notification channel. Run it for a few days to see if the incoming articles are indeed relevant.
After that, gradually add improvements:
- Change instant notifications to daily digests.
- Add automatic summaries with clear length limits.
- Keep logs of processed articles for easy auditing.
- Create an approval path for external messages.
- Review keywords based on actual results, not assumptions.
A good RSS to WhatsApp automation is not the one that sends the most messages. Its value is seen when recipients receive less information, but it is more relevant, easier to understand, and arrives at the right time.
– Rio Yotto @rioyotto