Free n8n workflow · importable JSON

Automated email news watch

Every morning at 8am, this workflow reads your RSS feeds, keeps what was published in the last 24 hours, has an AI model write a digest — articles to read first, then the rest grouped by topic — and delivers it to your inbox. Same recipe as the FlowKit pack workflows: official nodes, a carefully written prompt, explanatory notes on the canvas.

$ n8n import:workflow --input=flowkit-email-watch-free.json
✔ Import successful — 8 nodes, valid connections

How to import it, step by step

1. 01

   Download the JSON file

   The button above downloads flowkit-email-watch-free.json — a native n8n export, the same format as the FlowKit packs.

2. 02

   Import it into n8n

   In n8n (Cloud or self-hosted, version 1.60+): Workflows → ⋯ menu → “Import from File”, then pick the file. The 8 nodes, their connections and the explanatory notes land as-is on your canvas.

3. 03

   Attach your OpenAI credential

   Open the “OpenAI model” node and select (or create) your OpenAI credential. The default model is gpt-4o-mini: expect a few cents per month.

4. 04

   Set your RSS feeds

   Open “Feeds to watch” and replace the two example feeds with your own sources: industry blogs, competitors, trade press, alerts.

5. 05

   Configure the email

   In “Send the digest by email”, set the sending address, yours as recipient, and your SMTP credential (your host's or a service like Brevo).

6. 06

   Activate the workflow

   “Active” toggle, top right. Tomorrow at 8am your first digest arrives. To test right away: “Execute workflow”.

What each node does

1. 01

   Every morning at 8am

   Schedule (cron)

   Triggers the workflow every morning at 8am (cron “0 8 * * *”). Switch to “0 8 * * 1” for a weekly Monday digest.

2. 02

   Feeds to watch

   Code

   Your RSS feed list, in plain sight in a small Code node: one URL per line, add as many as you like. Two example feeds are prefilled (n8n blog, Hacker News).

3. 03

   Read the RSS feeds

   RSS read

   Reads every feed in the list and returns all published articles with title, link, excerpt and date. Works with any public RSS or Atom feed.

4. 04

   Keep the last 24 hours

   Code

   Keeps articles published since yesterday, deduplicates by link, strips HTML from excerpts and caps at 30 articles to control token costs. If nothing is left, the workflow stops: no empty email.

5. 05

   Aggregate the articles

   Aggregate

   Bundles all articles into a single JSON item, passed to the model in one call — one LLM call per day, not one per article.

6. 06

   Write the digest

   LLM chain

   The LLM chain applies a complete news-watch prompt: 3 to 5 “read first” articles with one sentence of context each, then the rest grouped by topic — clickable links, never an invented article.

7. 07

   OpenAI model

   OpenAI model

   gpt-4o-mini by default: a daily digest costs about a cent. Swappable for an Anthropic (Claude) node without touching the rest of the workflow.

8. 08

   Send the digest by email

   Send email

   Sends the HTML digest through your SMTP (Gmail, Brevo, your host…). Set your from/to addresses, attach your SMTP credential, done.

Customisation ideas

• →Weekly instead of daily: cron “0 8 * * 1” and a 168-hour window (one constant to change, flagged in the canvas notes).
• →Digest on Slack or Telegram alongside email: add a Slack or Telegram node after “Write the digest”, as in the AI Inbox Pack.
• →Switch to Claude: replace the “OpenAI model” node with an Anthropic node — the LLM chain stays the same.
• →Focus the digest on your topics: add your priority keywords to the prompt in “Write the digest”.