News Roundup – News roundup prompt for Perplexity, ChatGPT and Gemini
Free with your account
- Free with an account
- Copy filled in or download
- Updates included
Claude Opus 5.5
This news roundup prompt collects what happened on a topic over a period you set, with dates and sources. Add the topic, the period and the reader; you get the key developments, why each matters and what to watch next.
What you get
- A two-sentence headline on the biggest change in the period
- 5 to 8 dated developments, each with why it matters to your reader
- Rumours and single-source reports labelled Unconfirmed
- Coming dates and decisions to watch, with sources
How to use it
- 1Set the period and the reader
Give a period such as the last 7 or 30 days, and say who reads the roundup, so it keeps only what matters to them.
- 2Run it with web search
Use Perplexity or a research mode. A model without search knows nothing after its training data and can only give background.
- 3Check the top items
Open the original announcement behind the top three items and confirm each date before you pass the roundup on.
Better results
- Save the prompt with your topic and reader filled in, and change only the period each week.
- Ask a follow-up: "Which of these items has a primary source, and which only press coverage?"
Questions
Can Perplexity summarize news?
Yes, Perplexity can summarise news on a topic from current web pages and link each point to its source. Set a period, name the reader and ask for a date on every item, then open the original announcement behind anything you pass on.
Can ChatGPT search the web for recent news?
Yes, ChatGPT can search the web when search is on, and it then cites the pages it used. Without search it answers from training data that ends at a fixed point, so for news, check that search is on and ask for a date on each item.
What is the best AI for news summaries?
The best AI for news summaries is one with live web search that shows its sources and dates, such as Perplexity, or ChatGPT, Gemini or Claude with search on. A model without search only knows events up to the end of its training data.
Updated 6 Oct 2026
























