Get started

AI web search with cited sources

A plain language model answers from its training data and can be confidently wrong about dates, prices and news. In Rapir you switch search on and the answer is assembled from current pages, with a link next to every claim so you can check it.

Try for free Starter credits, no card, no watermarks
Sources
A link on every fact
Freshness
News and articles as of the moment you ask
Mode
One button in the chat

How it works

  1. 1

    Turn search on

    Press the search button next to the input — it applies to the next message.

  2. 2

    Ask

    Ask as usual: “what happened this week in…”, “compare prices for…”, “find studies on…”.

  3. 3

    Check the sources

    Facts carry footnotes with links. Open the preview or go to the page.

Benefits

  • Current data instead of the model’s stale knowledge
  • A source link on every claim — easy to verify
  • A digest of several articles in one request
  • Works in the same chat as regular questions, documents and generation

When to turn search on

Search matters where freshness or verifiability matters: news and events, prices and availability, software and model versions, laws and rules, statistics, quotes and studies. For creative work, code and pure reasoning, leave it off — the answer is faster and cheaper.

  • Industry news for the week — short, dated, linked.
  • Comparing services or products from public sources.
  • Meeting prep: who said what, and when.

How to phrase the request

Name the period (“in the last month”), the region and the answer format (“a table”, “five bullets”). Ask for links explicitly when you need primary sources. Follow-ups go in the same thread — the context and the pages found are kept.

FAQ

From public web pages found for your query at answer time. The list of sources is shown next to the answer.

See also

Ready to try?

Open Rapir and get it done in a single chat — with starter credits and no watermarks.

Start for free