Auto-generate AI-summarized newsletter drafts from RSS with n8n

RSS feed → Information Extractor summary → OpenAI newsletter text → Gmail draft

Who

Operators or automation engineers who need to curate industry news into newsletters regularly

You get

Automatically fetch new RSS articles every hour, summarize and personalize them via AI, then save as Gmail drafts for human review before sending

You need

n8n · gmail

Steps

1. Configure RSS Feed Trigger

   Add an RSS Feed Trigger node in n8n. Set feedUrl to the RSS feed you want to monitor (default in template: https://www.artificialintelligence-news.com/feed/). Set pollTimes to your preferred schedule; the template uses everyHour, but you can change it to daily or a custom interval.

2. Configure Information Extractor for summarization

   Add an Information Extractor node and connect it to the RSS Feed Trigger. In the text field, enter the following expression to pass the RSS item title, snippet, and full content:

   Execution prompt · Source text。Information Extractor node text field

   Copy prompt

   =Snippet: {{ $json['content:encodedSnippet'] }}
   Title: {{ $json.title }}
   Content: {{ $json['content:encoded'] }}

3. Set extractor system prompt and connect language model

   In Information Extractor options.systemPromptTemplate, set the system prompt that defines output language and summary style. Then connect an OpenAI Chat Model sub-node (lmChatOpenAi) to the Information Extractor ai_languageModel socket; choose gpt-4.1-mini, gpt-4.1, or gpt-4.

   Execution prompt · Source text。Information Extractor → options → systemPromptTemplate

   Copy prompt

   Your task is a get information from content snippet, and translate it in {your prefer} language. 
   Keep it clear and short about what is in content and title.

4. Configure Message a model to generate newsletter

   Add a Message a model (openAi) node receiving output from Information Extractor. In messages, configure system and assistant roles: the system instruction references extracted data and uses a Patrik assistant persona to create newsletter content; the assistant instruction holds your desired tone. Model options: gpt-4.1-mini, gpt-4.1, or gpt-4.

   Execution prompt · Source text。Message a model → messages → system field

   Copy prompt

   Your task is get information about article, use assistant role named Patrik to give personalisation and create a newsletter content.
   Content si here:
   {{ $json.output.Content }}

   Execution prompt · Source text。Message a model → messages → assistant field

   Copy prompt

   Here give your tone of voice!

5. Configure Gmail Create a draft

   Add a Gmail node, set resource to draft and emailType to html. Keep subject as Newsletter or customize it. Set message to {{ $json.message.content }} to receive content from the previous step. Connect your gmailOAuth2 credential, choose sender and draft folder.

6. Activate and test the workflow

   Save and activate the workflow. Run a manual test first to verify that a properly formatted HTML draft appears in your Gmail drafts. Once confirmed, keep it active to run on the scheduled interval.

You are setting up an n8n workflow: fetch articles from an RSS feed, extract and summarize them with an AI model, generate newsletter text using a specified persona, and save the result as a Gmail draft. First confirm you have a working n8n instance, openAiApi credentials, and gmailOAuth2 credentials, plus your RSS feed URL. Then configure RSS Feed Trigger, Information Extractor (including systemPromptTemplate), OpenAI Chat Model sub-node, Message a model (with system/assistant messages), and Gmail Create a draft. Test end-to-end with sample data before enabling the schedule.