Zapier is a tool that allows users to automate workflows. Instead of manually entering data, it can connect an RSS feed to ChatGPT and automatically analyze or summarize new content as it becomes available.
Prerequisites
Step 1: Create an RSS Feed with RSS.app
1. Copy the URL of any website, social media page, or other online source you want to track.
2. Paste the URL into our RSS Generator.
If the site is highly customizable or you just want more control over what to include in the feed, try using our RSS Builder.
3. Click Generate.
4. Your RSS feed should be ready within 20 seconds. Click Save to my Feeds to save it.
Use filters to customize the content that appears in your feed. You can exclude/include posts based on keywords, add custom text, replace text, etc.
Step 2: Connect your RSS feed to Zapier
1. Log into zapier.com.
2. Click Create New Zap.
3. For the Trigger, search for RSS and select it.
4. Choose New Item in Feed as the trigger event and click Continue.
5. Paste in the feed URL from RSS.app and click Continue.
6. Click Test Trigger.
You should multiple records being pulled from your RSS feed. Choose one to use as a test with the rest of the Zap.
Step 3: Generate your OpenAI API keys
1. Log into https://platform.openai.com/.
2. In the left sidebar, click API Keys.
3. Click + Create new secret key.
4. Give your key a name like “Make.com RSS Summarizer,” then click Create secret key.
5. Copy the resulting key.
Step 4: Connect Zapier to ChatGPT
1. Search for and select ChatGPT (Open AI) as the action in your Zap.
2. Choose Send Prompt as the action event and click Continue.
3. In the Model dropdown, select the AI model you want to use.
4. Type in your instructions in the Prompt field. To give the AI access to your RSS data, type / and choose the fields from your RSS step.
5. The Temperature controls how creative or deterministic the AI should be:
Lower values (e.g. 0–0.3) → more factual, consistent wording.
Medium values (e.g. 0.5–0.7) → balanced, natural writing.
Higher values (e.g. 0.8–1.0) → more creative and varied, but less predictable.
6. The Maximum Length sets the upper limit for how long the AI’s response can be (measured in tokens, roughly similar to words plus punctuation).
Short outputs (tweets, captions): 50–150
Medium outputs (summaries, LinkedIn posts): 200–300
Longer outputs (email drafts, longer summaries): 400+
If you are not sure, set this to around 300 to prevent overly long responses.
7. Configure additional options (optional):
Stop Sequences – Tells the AI where to stop generating text when it encounters a specific sequence of characters. Leave blank unless you need the output to stop at a custom marker.
Top P – Another control for randomness, similar to Temperature. A value of 1 is standard and works for most use cases.
Frequency Penalty – Discourages the AI from repeating the same phrases too often. Default 0 is fine for most setups.
Presence Penalty – Encourages the AI to introduce new topics or phrases. Default 0 is also fine.
If you do not have a specific reason to change them, you can keep all of these at their default values and move on.
7. Click Test Step.
You should see the ChatGPT-generated answer in the Data Out tab.
Step 6: Route or Inspect the Summary
You can now either:
Email yourself: Add an “Email” module and map the summary into the body to receive a digest by email.
Save somewhere else: Connect to Google Sheets, Airtable, Slack, etc.
For a quick test, you can do a Test Run once and inspect the ChatGPT response in Zapier. It will show you exactly what text was returned.