Build a daily AI news briefing with n8n and Claude

The Schedule Trigger starts the briefing at the same time every morning, with no clicks from you.

1. 1In n8n, create a new workflow and add a Schedule Trigger node.
2. 2Set Trigger Interval to days, Days Between Triggers to 1 and Trigger at Hour to 7am.
3. 3Open the three-dot workflow menu, pick Settings and set Timezone to your own.

You should see a Schedule Trigger node on the canvas, set to run once a day at 7am.

RSS feeds give clean titles and links, and a cap of 5 stories per feed keeps the prompt short.

1. 1Add an RSS Read node after the trigger and paste the TechCrunch AI feed below into URL.
2. 2Add a second RSS Read node after the trigger, with the Hacker News feed below.
3. 3After each RSS Read node, add a Limit node with Max Items set to 5.
4. 4Connect both Limit nodes to a Merge node, with Mode set to Append.

https://techcrunch.com/category/artificial-intelligence/feed/
https://news.ycombinator.com/rss

You should see about 10 stories in the Merge node's output, each with a title and a link.

To test one node while you build, open it and click Execute step. n8n runs the nodes before it too.

Claude should get one clean list in a single call, not one call per story.

1. 1After the Merge node, add a Filter node. Add 3 String conditions on {{ $json.title }}, each set to does not contain: Sponsored, Advertisement and Partner.
2. 2Combine the conditions with AND, and turn on Ignore Case.
3. 3Add an Aggregate node and pick All Item Data (Into a Single List).
4. 4Set Put Output in Field to stories. Under Include, pick Specified Fields and list title, link and contentSnippet.

You should see 1 item in the Aggregate output, holding a stories list.

Claude turns the raw list into a short briefing and picks the one story worth your time.

1. 1Add a Basic LLM Chain node after the Aggregate node. Set the prompt source to Define below and paste the prompt below into Prompt (User Message).
2. 2Connect an Anthropic Chat Model node to the chain's model input.
3. 3For its credential, paste an API key from platform.claude.com/settings/keys. On n8n Cloud, you can pick Use Gateway credits instead.
4. 4In Model, choose a fast model from the list, such as Claude Haiku 5.5.

You write a short morning briefing on AI news for a busy reader.

For each story below, give the headline on its own line, then 2 sentences on what happened and why it matters, then the link. After the list, name the 1 story most worth reading today and say why in 1 sentence. Write plain text that reads well in Slack, with no tables, no JSON and no markdown headings.

Stories:
{{ JSON.stringify($json.stories) }}

You should see the briefing in the chain's output, in a field called text, with a pick of the day at the end.

Your API key lives in an n8n credential, never in the prompt. Keep it out of shared files, exported workflows and screenshots.

A test run catches formatting problems before the first 7am run, and publishing turns the schedule on.

1. 1Add a Slack node with the message Send operation. Set Send Message To to Channel, pick your channel and set Message Text to {{ $json.text }}.
2. 2For the Slack credential on n8n Cloud, click Connect my account and sign in to Slack.
3. 3Click Execute workflow and check the message in Slack.
4. 4Click Publish in the canvas header, then Publish again in the dialog.

You should see the briefing in your Slack channel, and the workflow marked as published.

A Schedule Trigger only fires in a published workflow. Changes you make later stay in draft until you publish again.