Most AI reads arented search result.This one reads the web.

Four things that are true here and rare elsewhere.

None of these is a claim about being biggest or fastest. They are claims about shape. A library is not useful because it owns the most paper; it is useful because someone shelved the law reports with the law reports, and will hand you the page rather than the catalogue. Each of these is checkable on a page of this site.

Not our opinion. The record.

Owning the index used to be an eccentric thing to do. Over two years the case for it stopped being an argument and became a series of events, none of which we had any hand in. Here they are, most recent first, with the source for each — move along the track to read back through them. Only the two at the front are ours, and they are marked.

The entry worth sitting with is the first one, because of who wrote it. In 2024 the biggest AI product on the web launched by leveraging third-party search providers, and said so plainly. Everything after it is the market finding out what that dependency costs — an endpoint retired at short notice, a lawsuit over relayed results, two independent indexes going dark. The conclusion is not ours to argue; it is what the record already says. What that looks like here is 30 collections you can open and read.

Many small machines, not one rented service.

Reading the web continuously is the expensive part of running an index, and it is the reason most products rent their results instead. askFinz spreads the work across pools of ordinary machines — a spare desktop, a container on a server already running, or a dedicated box booted from an image — each joining a pool and taking work when it has capacity.

The node programme is open — lend a machine, run it as a container , or boot a dedicated box from an image. Why we do not use a rented proxy pool is set out in how askFinz reaches a site.

At the floor, and honest about it.

Every rate below was read off the vendor’s own pricing page, and every capability off their own documentation, on 7 September 2026. The table is grouped by the thing that matters most and is hardest to see from a rate card: whether a provider holds an index of its own, borrows one, relays a search engine, fetches a page you already named, hands you an archive to process, or meters the web inside a model. Cheaper rows are in here rather than hidden underneath — most of them sell a call that never returns the words on the page, which is a column now instead of a footnote.

Capabilities from each vendor’s own documentation, rates from their own pricing page, both checked 7 September 2026. Rates are the cheapest published self-serve figure per 1,000; where a vendor publishes none, the cell says what they price by instead rather than carrying an estimate. Free tiers are excluded throughout — they are a trial, not a price. Rows run cheapest first within each group.

Provider	What comes back	Filed by document type	Workspaces on top	From, per 1,000	vs askFinz bundled
Holds its own indexRuns its own crawler and answers from what it holds — a library with its own shelves. The only group that decides what is on them.
askFinz	The passage that answered you, and the address to check it against — as many as the question needsIncluded	Shelved by what it is: a judgment comes back as a judgment, with its citation. 45+ collections	Open it in Search, keep it in Research, work in a browser extension or the desktop app	$0.75top-up; $1.00 with the extension or a paid plan, $3.50 standalone	—
PA Parallel	Ranked links, with shortened extractsIncludedextracts, not the full page	—Ranked web pages	—An API endpoint	$1.00turbo and fast; covers the first 10 results, then $1.00 per 1,000 further results. Basic and advanced $5.00	about the sameat 10 results a call. A call returning 100 costs about $91 per 1,000 on the same published rates
KE Keiro	Ranked links. Page text costs more creditsIncluded3 credits; links alone cost 1	—Ranked web pages	—An API endpoint	$2.40best tier; $4.00 and $7.20 below it. Links only $0.80	2.4× more
KE Keenable	Ranked links, with long snippetsBilled extrasnippets, not the page	—Ranked web pages	—An API endpoint	$4.00pay as you go; $1.00 only on dedicated capacity	4.0× more
YE Yep	Ranked links, plus the sentence that matchedBilled extrathe matching sentence only	—Ranked web pages	—An API endpoint	$4.00advanced tier $8.00	4.0× more
BS Brave Search API	Ranked links, each with a descriptionNot returned	—Ranked web pages	—An API endpoint	$5.00answers tier $4.00 plus tokens	5.0× more
EX Exa	Ranked links. Page text is billed per pageBilled extra$1.00 per 1,000 pages on top	—Ranked pages, matched by vector similarity	—An API endpoint	$5.00answer tier; search $7.00 for 10 results, then $1.00 per 1,000 more; deep $12.00, deep-reasoning $15.00	5.0× more
LI Linkup	Links with content, or a composed answerIncluded	—Ranked web pages	—An API endpoint	$5.00standard depth; deep search roughly ten times this	5.0× more
AB Amazon Bedrock	Passages, returned inside BedrockIncluded	—Ranked web pages	—Inside Bedrock, with the rest of AWS	$7.00Amazon's own index, launched August 2026	7.0× more
Partly its ownRuns some crawling of its own and leans on an outside engine for the rest.
YO You.com	Ranked links. Page content is billed per pageBilled extracontents $1.00 per 1,000 pages	—Ranked web pages	—An endpoint, plus a consumer chat app	$5.001–100 results per call	5.0× more
PE Perplexity	Ranked links, with filtersNot returned	—Ranked web pages	—An endpoint, plus a consumer chat app	$5.00Search API, per 1,000 requests	5.0× more
GV Google Vertex AI Search	Results from Google's index, with groundingIncluded	—Your own data store, plus Google’s index	—A cloud console you configure	not per searchper query PLUS per GB indexed — no single per-1,000 figure	—
Relays a search engineHolds no index. Passes your query to a search engine and returns what came back. Cheaper, and a different product.
DA DataForSEO	A search engine's result page. No page textNot returned	—The engine's ranked links	—An API endpoint	$0.60queued mode; live calls $2.00	cheaper, thinner call
SE SearchAPI	A search engine's results — no page textNot returned	—The engine's ranked links	—An API endpoint	$1.00at volume; $4.00 at the entry plan	about the same
BD Bright Data	A search engine's result pageNot returned	—The engine's ranked links	—An API endpoint	$1.30SERP API; $1.50 without a commitment	1.3× more
SE SerpApi	A search engine's results — no page textNot returned	—The engine's ranked links	—An API endpoint	$1.97at the largest published tier; $25.00 entry	2.0× more
TA Tavily	Sources gathered from other engines, with content and a summaryIncluded	—Sources from other engines, behind one summary	—An API endpoint	$5.00largest plan; $8.00 pay as you go, advanced double	5.0× more
SE Serper	A search engine's results — no page textNot returned	—The engine's ranked links	—An API endpoint	not per searchno public per-1,000 rate published	—
Metered by what you consume, not per searchSold as a pool you spend on whatever you do with it — reading a URL you already have, running a scraper, or a search — so there is no single per-search rate to set beside the rows above. Both publish their arithmetic; neither publishes where a search result comes from.
FI Firecrawl	One URL you name, as markdown. Search is billed from the same creditsBilled extraa search costs 2 credits, scraping each result costs 1 more	—The page you asked for, or ranked links	—An API endpoint	not per searchcredits, not calls — about EUR 5.60 per 1,000 searches on the entry plan and EUR 1.14 at the top one, or EUR 2.80 and EUR 0.57 per 1,000 pages. Published in euros, so no dollar figure here	—
AP Apify	Whatever the scraper you picked collectsIncluded	—Whatever the scraper you chose collects	—A marketplace of scrapers you run	not per searchpriced per run and per result, not per search	—
A bulk archive you process yourselfNo queries and no ranking. A snapshot of the web delivered as files, which you store, parse and search on your own hardware.
CC Common Crawl	A monthly archive of raw pages. You process itIncludedraw, unparsed, months old	—Raw archive, unsorted	—Files you download	not per searchfree to download; you pay in storage and compute	—
Web access billed inside a modelNot an index you can query — the model's own access to one, metered per call by whoever runs the model, with the pages it read arriving as tokens you also pay for. Read the multiplier here as a rough guide only: it sets a per-call tool beside an index subscription, which is a different shape of purchase. Worth noting anyway that five companies publish a figure for the same verb and the spread runs six times end to end.
XW xAI web search	Searches, and opens pages, for the modelBilled extraarrives as tokens in the completion	—Whatever its search returned	—Inside that model's own product	$5.00per 1,000 web_search tool calls — the cheapest published figure in this group	5.0× more
OW OpenAI web search	Results passed to the model, inside the completionBilled extrasearch content tokens billed at model rates	—Whatever its search returned	—Inside that model's own product	$10.00web search, and the preview tool on reasoning models including gpt-5 and the o-series. $25.00 on non-reasoning models, where search content tokens are free	10.0× more
AW Anthropic web search	Results handed to the model, inside the completionBilled extracharged on top of token costs	—Whatever its search returned	—Inside that model's own product	$10.00per 1,000 searches through the API, on top of token costs	10.0× moreand US-only data residency via inference_geo adds 10% to all Opus 4.7 and above token costs, which base model pricing does not show
GG Google grounding	A grounded answer, not an index you can queryBilled extraarrives as tokens in the completion	—Whatever its search returned	—Inside that model's own product	$14.00per 1,000 grounded queries	14.0× more
MW Mistral web search	Results with citations, inside the model's replyBilled extraarrives as tokens in the completion	—Whatever its search returned	—Inside that model's own product	$30.00per 1,000 calls	30.0× more

How to read this. Check how many results a rate covers before setting two side by side: much of this market sells the first ten and meters everything after, so a question that has to weigh forty sources can cost several times the advertised figure. Then read the middle three columns — half the cheapest rows return no page text at all, and only askFinz files what it reads by what the document actually is or puts workspaces on the index rather than handing back an endpoint. Several rows beat askFinz at something: Parallel matches the price at ten results a call, Firecrawl reads a named URL askFinz will not, the relays are far cheaper for ranked links, and Common Crawl costs nothing if you own the hardware to process it — each of those pages says so. Every figure is the vendor’s best published self-serve rate, read from their own pricing page on 7 September 2026, matched on what the call returns and excluding free tiers. Bing’s Search API, which much of this market was built on, was retired in August 2025; the rest reprice often enough that you should check theirs before relying on this.

Five questions that settle it faster than a rate card.

1. 01

   Do you need the page, or a link to it?

   The cheapest services hand back a search engine's ranked links. If your model then has to fetch and read each one, that fetch is the real cost — and it is not on the rate card you compared. Ask how many results the rate covers, too: a top-ten window is the norm, and a question that needs to weigh more than ten sources is priced accordingly.

   Compare the rates, and what each call returns
2. 02

   Does it matter what kind of document it is?

   For general questions, no. For a judgment, a standard, a filing or a clinical study it decides everything: those carry identifiers, jurisdiction and status that a generic page index throws away.

   See every collection and how it is filed
3. 03

   Who owns what you are searching?

   A provider relaying an engine can only reach what that engine chose to show, and their price moves when the engine's does. A provider with its own index sets both.

   Owning an index versus renting one
4. 04

   Is this a call in a loop, or work someone does?

   If you are building an agent framework, an API with good integrations may save more in engineering time than it costs in rate. If people are doing the work, a workspace they can open beats an endpoint.

   The workspaces built on the index
5. 05

   What happens when the answer is questioned?

   Ask whether you can get back to the source. An answer with the passage and its address survives being challenged; a summary of a summary does not.

   How the index works, end to end

Affordable to run is why it is affordable to buy.

The point of keeping it cheap to run is that the people who most need an index rarely have a budget for one — a researcher, a small firm, a team automating something nobody has automated before. A rate that tracks hardware rather than a rented plan is what makes building on the web affordable for them, which is the argument for owning the thing rather than renting it. Owning an index versus renting one sets that out in full.

When something else is the better buy.

Every comparison on this site is written to survive the other company reading it. Rates come from their own pages, capabilities from their own documentation, and where a rival is cheaper or does something askFinz does not, the page says so. The full set is at compare, and the two category explainers are owning an index versus renting one and indexing versus scraping.

The ones worth asking first.

Is askFinz the cheapest web index?

It is at the floor rather than below it, and the honest answer depends on what the call returns. Among services that hold their own index and return the matched passage, askFinz's bundled $1.00 per 1,000 is the cheapest published rate, tied with one other. Services that return ranked links without the page text are cheaper — that is a thinner call, and a fair comparison says so. The full table, with every rate read off the vendor's own page, is on the pricing page.

What makes this different from a search API?

A search API gives your code results to reason over. askFinz gives an answer with the passage it came from, on an index filed by document type, inside workspaces you can actually work in. If what you want is an endpoint to build against, several of the services compared here are a better fit and the comparison pages say which.

Does askFinz rent its search results from a search engine?

No. The index is read and held by askFinz. That is unusual but not unique — a small number of others run their own index too, and the comparison pages name them rather than implying otherwise.

How does the reading fleet keep costs down?

Reading is spread across pools of ordinary machines rather than concentrated in rented capacity billed per request. Adding capacity means adding a machine to a pool, so the cost of reading more of the web tracks hardware rather than someone else's rate card. That is the mechanism behind the published rate.

Can I use askFinz for crawling or scraping?

Not directly. Scraping and crawling services take a URL you name and return its content; askFinz answers a question from material it has already read and filed. If your job is to fetch specific pages on demand, the crawling comparison explains which tools do that and why askFinz is the wrong answer for it.