Skip to content

One email a day with the news on your topic, summarised

Name the topic and a few sources. Each morning ClickRuntime collects new items from those feeds, has AI pick the ones that matter and write a short summary, and emails you one digest.

Your assistantClickRuntime connected
Create a ClickRuntime automation: every day at 7:30, collect yesterday's items from these AI policy feeds, pick the 5 that matter most and email me a short digest.
I'll gather new items from the feeds each morning, ask the built-in AI to pick and summarise the top 5, and email the digest at 7:30.
AI policy digestLive
When
Every day at 7:30 (Europe/London)
Uses
4 feeds, Email to you
RunsClickRuntime dashboard
  • 7:30✓ AI policy digest · SucceededLast run7:30 · 23 items, 5 picked

Can ChatGPT or Claude email me a daily news digest on a topic?

They can summarise news when you ask, but they won't send it to your inbox every morning from sources you chose, and they don't remember what you were sent yesterday. With ClickRuntime connected, your assistant sets up an automation that collects new items from your feeds each day, uses the built-in AI to pick and summarise the best, and emails you the digest.

Why the assistant can’t do this on its own

  • A daily email needs a schedule and a way to send email outside the chat.
  • Your sources, not a search engine's choice: the automation reads exactly the feeds you list.
  • No repeats: it keeps the IDs of items already included, so yesterday's story doesn't come back.

Set it up from your AI assistant

The same steps work in ChatGPT, Claude, Cursor and any assistant that supports MCP connectors.

  1. Step 1Connect ClickRuntime to your assistantAdd ClickRuntime as a connector (MCP server) in ChatGPT, Claude, Cursor or another MCP client, using https://clickruntime.com/api/mcp. You sign in with Google or email; there is no API key to copy. This is done once per assistant.
  2. Step 2Paste the promptStart a new chat and paste: “Create a ClickRuntime automation: every day at 7:30 Europe/London, collect new items from these feeds: <feed URLs>. Pick the 5 most important for someone following AI policy in the EU, and email me a digest with a two-sentence summary and link for each. Don't include anything I was sent before.” Change the details (times, channels, addresses) to yours.
  3. Step 3Read the draftThe assistant restates what will happen, asks for anything missing, such as your timezone or which channel to post to, and saves a draft automation. A draft never runs on its own, so nothing happens until you agree.
  4. Step 4Connect accounts and save keysNothing to connect for public feeds. The digest is emailed to the address you signed in with.
  5. Step 5Test itThe assistant runs it once now and you get today's digest by email. It adjusts the topic description if the picks aren't what you want.
  6. Step 6Turn it onThe assistant tells you what will run, when, and which services it will contact, then asks your permission to turn it on. From then on it runs on ClickRuntime's servers, and you can close the chat.
  7. Step 7Check it in the ClickRuntime dashboardEach run shows how many items were collected and how many were picked. The run's result keeps the titles that were sent; state holds the IDs to skip next time.
Prompt to copy.txt

“Create a ClickRuntime automation: every day at 7:30 Europe/London, collect new items from these feeds: <feed URLs>. Pick the 5 most important for someone following AI policy in the EU, and email me a digest with a two-sentence summary and link for each. Don't include anything I was sent before.”

Connector setup for each assistant: ChatGPT, Claude, Claude Code, Cursor, and other assistants.

What the assistant writes

A typical script for this job. Yours matches your pages, accounts and wording, and you can read every version in the ClickRuntime dashboard.

daily_news_digest.pyexample
1import feedparser
2from bs4 import BeautifulSoup
3from clickruntime import on_trigger, state, notify, ai, http, log
4 
5FEEDS = [
6"https://example-policy-news.org/feed",
7"https://another-source.eu/rss",
8]
9TOPIC = "AI policy and regulation in the EU, for a product lead at a software company"
10SCHEMA = {"type": "object", "properties": {"picks": {"type": "array", "items": {
11"type": "object", "properties": {"n": {"type": "integer"}, "summary": {"type": "string"}},
12"required": ["n", "summary"]}}}, "required": ["picks"]}
13 
14@on_trigger
15def run(event):
16seen = set(state.get("seen") or [])
17items = []
18for url in FEEDS:
19for e in feedparser.parse(http.get(url).content).entries[:30]:
20key = e.get("id") or e.get("link")
21if key and key not in seen:
22text = BeautifulSoup(e.get("summary", ""), "lxml").get_text(" ")[:400]
23items.append({"key": key, "title": e.get("title", ""), "link": e.get("link", ""), "text": text})
24if not items:
25notify.email_me("Your digest: nothing new today", "None of your sources published anything new.")
26return {"items": 0}
27 
28listing = "\n".join(f"{i}. {it['title']} — {it['text']}" for i, it in enumerate(items[:60]))
29try:
30picks = ai.generate(f"Topic: {TOPIC}\n\nItems:\n{listing}",
31system="Pick up to 5 items that matter most for the topic. Summarise each in two plain sentences.",
32json_schema=SCHEMA, max_output_tokens=900).json["picks"]
33body = "\n\n".join(f"{items[p['n']]['title']}\n{p['summary']}\n{items[p['n']]['link']}"
34for p in picks if 0 <= p["n"] < len(items))
35except ai.AIError as e:
36log.warning("AI unavailable, sending titles only", {"code": getattr(e, "code", None)})
37body = "\n\n".join(f"{it['title']}\n{it['link']}" for it in items[:10])
38 
39notify.email_me("Your daily digest", body)
40state.set("seen", list(seen | {it["key"] for it in items})[-3000:])
41return {"items": len(items)}
  • One AI call per day: the items go in as a numbered list and come back as picks with summaries, using json_schema.
  • Only titles and a short excerpt are sent to AI, which keeps token use low.
  • If AI is unavailable, you still get an email with the titles.

A schedule, not a trigger link

A daily schedule in your timezone. For a weekly digest, ask for “Mondays at 8” instead.

Limits

  • Sources need to be RSS or Atom feeds, or sites with an API. For a site without a feed, the assistant can read its news page.
  • notify.email_me sends to you only. To send the digest to a list, the assistant uses an email service such as Resend or Mailchimp with your key.
  • Each AI call reads up to about 32,000 tokens, so very many sources are trimmed to the newest items.

Troubleshooting

The picks aren't what I care about.

Describe your interest more precisely to your assistant, with an example of a good and a bad item. It updates the topic text.

A feed fails to load.

The run log names it. Ask the assistant to check the feed URL or drop it.

What you can add next

The first version is rarely the last. Because the automation is a Python script with memory, AI and any API available, you grow it by asking your assistant for the change. Each change is a new version, and a failed deploy never replaces the working one.

  • Any app or API

    Send the digest to your team through Resend

    “Change my news digest so it's also sent to policy-team@example.com through Resend.”
  • Built-in AI

    Add one line on what each story means for your company

    “Change my news digest so each pick ends with one line on what it could mean for a company like ours.”
  • Memory between runs

    Follow a story over several days instead of repeating it

    “Change my news digest so if a new item is about a story I was already sent, it's grouped as an update to that story.”
  • Any app or API

    Save picks to a Notion database you can search later

    “Add to my news digest: save each pick to my Notion database News with date, title, link and summary.”
  • Schedules

    A longer weekly round-up on Fridays

    “Add to my news digest: every Friday at 16:00, email me a round-up of the week's 10 most important picks.”
  • Multi-step logic

    Send urgent items straight away instead of waiting for tomorrow

    “Change my news digest so it also checks hourly and emails me at once if an item mentions a new law being passed.”

Can ChatGPT email me a news summary every day?

With ClickRuntime connected, yes. ChatGPT sets up an automation that reads your sources every morning, summarises them with AI and emails you, after the chat is closed.

Can Claude send me a daily briefing on a topic?

Yes, through ClickRuntime. Claude writes it; ClickRuntime runs it on schedule.

Do I need an OpenAI or Anthropic key for the summaries?

No. Automations use ClickRuntime's built-in AI, including on the free plan.

Will it repeat stories?

No. It keeps the IDs of items already sent.

Can I choose the sources?

Yes. You list the feeds or sites; the automation reads only those.

What can I add later?

Team delivery, impact notes, story tracking or a Notion archive. See Extend it.

Other use cases

Set up your first automation in ChatGPT or Claude

Free plan. No credit card.

Server URLhttps://clickruntime.com/api/mcp