A free web search engine with its own index, no ads, no account and no third-party scripts, that ranks clean pages above ad-heavy ones. Here's what it records.
Carpathian Search is a free web search engine at https://search.carpathian.ai. It runs its own crawler and its own index, shows no ads or sponsored results, and needs no account. It ranks pages by how well they match your search first, then, among pages that match about as well, lifts the ones without ad networks, popups, cookie walls and trackers.
It's a private search engine in specific, checkable ways: it never stores your IP address or user agent, loads no third-party scripts, and sells nothing to advertisers. It does record what's searched, to measure and improve results, and the privacy section below lists exactly what. There's also a Search API for developers and AI agents, free up to a monthly allowance.
How is it different from the search engines you use now?
The results come from Carpathian's own index, built by its crawler, CarpathianBot, not resold from a larger engine. That changes what you see in three ways.
No ads or sponsored placements. Every result is there because it matched your search.
Clean pages rank higher. Among pages of similar relevance, pages without ad networks, popups covering the content, cookie walls, analytics trackers and affiliate listicles rank above pages that have them. A recipe delivered as text ranks above the same recipe wrapped in a megabyte of scripts.
Opinion searches favor people. A search with words like best, vs, review or worth it gives more weight to first-hand accounts and discussion threads, and a stronger penalty to affiliate content.
These signals never lift an irrelevant page above a relevant one. They only reorder pages that match about equally well, and every result has a "Why here" link showing the signals behind its place.
How private is Carpathian Search?
Carpathian Search keeps no account, stores no IP address or user agent, and loads no third-party resources. It does record searches and clicks to tune its ranking. Its privacy page lists everything, and the outline is below.
What it never stores:
IP addresses and user agent strings. Your address is read to look up a rough location in a database on the server itself, and your user agent to read the device type, then both are discarded.
Searches that hold personal details. A search containing an email address, a phone or card number, or a URL with parameters is counted and its text thrown away.
Searches by bots and scripts. They're answered but not counted.
What it records:
Each search and click: the query, the results shown, which result was clicked and how long until you came back.
Coarse context: country, region and city looked up from your address in a local database, device type, browser and operating system family, and your browser's language.
Visits: whether you're new or returning, grouped into sessions.
Daily totals are kept for 400 days. A log of individual searches and clicks is kept on the search server for 90 days to improve ranking, then deleted. Neither is sold or shared with advertisers.
Does it use cookies?
The site sets one first-party cookie, cs_vid, holding a random identifier so a returning visit can be recognized. It expires 30 days after your last search, and scripts and other sites can't read it.
In the European Economic Area, the United Kingdom and Switzerland, and wherever your location can't be determined, it's set only if you choose Accept.
If your browser sends Global Privacy Control, it's never set, and one already set is deleted.
Without the cookie, your searches are grouped under a temporary id made from a keyed hash that's replaced every day and never written to disk, so it can't be traced back to your address or joined across days.
How do you search with it?
Type what you're looking for. Search operators narrow it down:
"cast iron" finds the exact phrase.
skillet -nonstick leaves out pages with a word.
skillet site:example.com searches one site and its subdomains.
Case, accents and simple plurals don't matter, and a misspelled word is corrected, with a link back to what you typed. The full list is on the search operators page.
Can your site appear in it?
Yes, and it costs nothing. CarpathianBot follows robots.txt and every opt-out on the opt-out page. To get a site crawled sooner, submit it or its sitemap on the sitemap page, or tell the engine which pages changed with IndexNow.
Site owners can add their site to Search Engine Console in Carpathian at no charge, prove they own it with a DNS record or a meta tag, and see the searches their pages appear for, with impressions, clicks and average position. They never see who searched.
Can developers and AI agents use it?
The Search API returns the same results as JSON, and reads the stored text of any indexed page. It's built for applications and AI agents that need to search the web, with tool definitions for function calling in its documentation. Make a key in Search Engine Console, and searches and page reads are free up to a monthly allowance, then billed per 1,000.
The index is Carpathian's own, so its coverage differs from the engines you're used to, so try your usual searches before you switch. It returns web pages, not image, video or shopping results. And it isn't a no-logging search engine: queries and coarse location are recorded, as listed above, to make the results better.
Try a search
Search at https://search.carpathian.ai, no account needed. If you run a site, create an account and add it to Search Engine Console to see the searches it shows up for, and the same account makes Search API keys for your apps and agents.
Free Search Engine With No Ads Or Trackers | Carpathian