User Agreement

You type a question. We send that question to the services that find and read academic sources, and to the model that writes the answer. We don't sell your searches, we don't build an advertising profile on you, and we don't ask you to make an account.

The text of your question, and basic technical information that any website receives when you visit it, like your browser type, rough location from your IP address, and the time of the request. We don't ask for your name, email or payment details, so we don't have them.

To answer a question, your search text is sent to the specific third parties we rely on to run the search:

• Firecrawl, the web search and page-reading service that finds the academic sources and pulls their text. It receives your question and the pages we ask it to read.
• Lovable, which hosts this site and runs the AI Gateway that our requests pass through. It receives your question as part of that request.
• OpenAI, whose GPT models (currently GPT-5.6) evaluate the sources and write the answer, through the Lovable AI Gateway. They receive your question and the text of the sources we found. We ask for no training retention on these requests.

Each of these providers processes the text under its own terms, and we only send what is needed to run the search. If we ever add or swap a provider, this list gets updated.

We also fetch pages from the universities, journals, repositories and public-interest bodies in our source list, such as Harvard, Stanford, MIT, PubMed, arXiv, the WHO, the CDC, the NIH and Pew. Those sites see a request from our servers, not from you.

Searches are processed live and are not stored in a user profile. Short-lived operational logs may hold a request for a limited period so we can debug problems, catch abuse, and keep the service running. They are not used to identify you or to target you.

We don't sell or rent your searches. We don't run ads against them. We don't share them with data brokers. We don't use them to figure out who you are.

Use TheWorldSearch for research and learning. Don't use it to break the law, to scrape or overload the service, or to try to get around the sourcing rules. We can limit or block use that damages the service or the sites we read from.

Every answer is built from cited academic sources, and we refuse to answer when the sourcing isn't strong enough. That still isn't medical, legal or financial advice. Research changes, summaries lose nuance, and models make mistakes. Read the sources we show you before you act on anything important.

If this agreement changes in a meaningful way, we'll update the date at the top of this page. Questions about how your searches are handled can go to Nicholas Tropiano, who built and runs TheWorldSearch.