Best Pixel Position API Providers in 2026 (Who Actually Returns SERP Geometry)

Most “best API provider” posts are price ladders, because every vendor on the page sells the same thing and the only open question is what it costs. Pixel position is not that market. Before you can compare a price you have to find a vendor that returns the field at all, and on 16 September 2026 I could find four. Eight other well-known SERP vendors publish no pixel-geometry field anywhere on the product or pricing pages I read. That is the finding; the prices come after it. Checked 2026-09-16.

I read every rate below off the vendor's own live pricing page on 16 September 2026 and linked each one inline. I ran no benchmark and I make no claim about anyone's speed or success rate — this is a capability, coverage and published-price comparison built from vendor documentation. Serpent is the API this site sells, and it appears here as one row among several. If you want the concept rather than the vendors, the explainer on what pixel position actually measures covers it and this post does not repeat it.

How I decided who returns pixel position

The inclusion test is deliberately narrow, because “visual SERP” marketing language is everywhere and a coordinate is not. To be in the table, a vendor's own product, pricing or documentation page has to name a coordinate, a rectangle or a pixel rank as a field its API returns. A screenshot endpoint does not count: an image is not a measurement, and turning one into coordinates is your problem, not the vendor's. A “SERP layout” feature in a dashboard does not count either, because this is a comparison of APIs.

• Entry rate and lowest published rate are separate columns. Quoting a vendor's volume floor against another vendor's entry price is the commonest dishonesty in this category. Both are shown, and so is what the lower rate actually requires you to spend.
• If a vendor does not publish a figure, the cell stays empty and the row says why. There is one of those below: the high-priority row of DataForSEO's own cost table was not legible on either of my two reads, so that cell is blank rather than estimated.
• Exclusions are findings, not omissions. The eight vendors that publish nothing get their own named table further down. Leaving them out silently under a heading with the word “best” in it would be selection bias.
• Absence of evidence is not proof of absence. Saying a vendor publishes no pixel-geometry field is a claim about its published pages on one date. It is not a claim that the company could not build one, and several of them plainly could.
• Currency is not silently normalised. One of the four prices in pounds and publishes no dollar rate. It stays in pounds, in its own column, with the reason written next to it.

Twelve vendors went into the check: the four below, plus Bright Data, SearchApi, Oxylabs, HasData, Scrapingdog, Serper, Zenserp and Traject Data. That is the whole population, and one third of it passed.

The four APIs that publish a pixel-geometry field

Here is the capability table first, because on this product the capability is the decision and the price is a rounding error next to it.

Sources read 2026-09-16: serpapi.com/pixel-position-api, dataforseo.com/help-center/serp-api-cost-explained, authoritas.com/keyword-ranking-api, and our own Pixel Position API page. Quoted strings above are the vendors' own words; everything outside quotation marks is my summary of what the page says. No API was called and no benchmark was run.

SerpApi: the most explicit page, the narrowest engine list

SerpApi's pixel position page is the clearest of the four about what arrives in the response. It says the feature adds x, y, width and height to organic results, local results, the answer box “and many other elements”, and that you switch it on by appending .json_with_pixel_position to the /search endpoint or by passing output=json_with_pixel_position. It is also the clearest about its own limits: the page names the APIs the feature is available for, and the list is two entries long, the Google Search API and the Google Ads API. There is no third engine on it.

What the page does not contain is a price of any kind — no surcharge, no extra credit cost, no note that the output parameter changes the billing. That absence is worth stating precisely, because the precise version and the sloppy version differ: SerpApi publishes no surcharge for pixel position. That is not the same sentence as “pixel position is free on SerpApi”, and I am not going to turn one into the other. The full plan ladder, which is what a pixel-position request is actually billed against, is broken down in the SerpApi pricing explainer.

DataForSEO: the only vendor that puts a number on the field

DataForSEO's help centre explains the parameter in one sentence: calculate_rectangles is designed for retrieving pixel rankings for each element in the SERP, and to obtain pixel dimensions of the results you set it to true. It also does something none of the other three do — it puts a price on the field, in a published table, on its own site. The cost table lists the parameter's cost as base price plus base price, which is a polite way of writing “twice”.

On a queued Google organic SERP DataForSEO charges $0.60 without the parameter and $1.20 per 1,000 SERPs with it; in Live mode it is $2.00 against $4.00 per 1,000 SERPs. That single fact is the most useful thing in this post: on at least one major vendor, the geometry is a paid parameter that doubles the price of the call, not a free field that comes along for the ride. If you are budgeting a rank-tracking run in the millions of tasks, that multiplier is the line item, not the base rate.

The same page also says that applying one of its keyword operators multiplies the base price per task by five, and the cost table shows that multiplier stacking with the rectangle charge. I have not printed a combined per-1,000 figure for that, and I have not printed a priority-queue figure either: the high-priority row of that table came back malformed on both of my reads, and an unreadable cell stays empty here rather than becoming an estimate. The three queue modes and what they cost without the parameter are unpacked in the DataForSEO pricing explainer.

Authoritas: the broadest claim, the least specification

Authoritas is the vendor most SERP-API comparisons miss, because it is sold as a rank-tracking platform rather than as a scraping API, and it does not appear in the usual “SERP API alternatives” lists at all. Its keyword ranking API page makes the geometry claim in a single sentence: “We can tell you the (x,y) position of all elements on the page”. That is the broadest capability statement of the four — all elements, no feature list, no exceptions.

It is also the least specific. The page does not name a field, a parameter, a response shape, or the engines the coordinates are guaranteed on. It lists Google plus Bing, Yahoo, Yandex, Baidu and Naver as engines the API covers, but it does not tie the coordinate claim to that list, so I have not either. If you are evaluating Authoritas for this specifically, the question to put to their sales team is not “do you have pixel position” — the page already says yes — but “which engines, which elements, and what is the field called”.

Serpent: two fields, four engines, no surcharge, and a real gap

Ours is the fourth, and since this site sells it, here is the specification rather than the pitch. The parameter is pixel_position=true on GET /api/search, and every item the measurement matched comes back with two sibling fields: pixel_position, an integer y-coordinate measured from the top of the rendered page, and pixel_box, an object of {x, y, w, h} in the same coordinate space, where pixel_box.y equals pixel_position by construction. The fields appear on organic rows and on the AI Overview, featured snippet, knowledge panel, People Also Ask, top ads, local pack, inline video, shopping and related-searches blocks whenever the results page contains them.

Four engines carry the fields — google, bing, yahoo and ddg — and engine=brave does not. A Brave request with the flag set still succeeds and simply has no pixel fields on any item, which is exactly the sort of silent difference that breaks code assuming a key is present. Three more limits, stated plainly because you will hit them: it is BETA, it is desktop-only at a 1366×768 viewport, and a call served out of the free allowance returns no pixel fields at all. If your reporting is mobile-first, that last-but-one point is a real gap and you should know it before you integrate rather than after.

A coordinate and a rectangle are not the same product

“Returns pixel position” looks like a yes/no column and is not one. A single y-coordinate tells you where an element starts. A rectangle tells you where it starts, how tall it is, how wide it is and which column it sits in — and half the questions people actually buy this data to answer need the height, not the offset.

Read this against the capability table above, checked 2026-09-16. SerpApi and DataForSEO both publish rectangles, so both answer every row. Authoritas publishes an (x,y) pair and says nothing about size, so on its page the height rows are unanswered. Serpent publishes both shapes, with the single integer equal to the rectangle's top edge, so there is nothing to choose between them.

The practical consequence: if all you ever do is chart a number over time, a y-coordinate is the whole product and the extra fields are noise. If you are doing share-of-voice work — how much of the first screen belongs to whom — a y-coordinate on its own cannot answer a single question you have, and a vendor that ships only the offset has sold you half a feature. Our own walkthrough of that calculation is in pixel position versus SERP features, and the case for tracking both the ordinal and the coordinate is in pixel position versus rank.

The eight vendors that publish none of it

This is the part of the comparison that took the longest and is worth the most. Eight of the twelve vendors on the shortlist — including several that are household names in this category and that sell perfectly good SERP data — publish no pixel-geometry field on any product or pricing page I read. Not a coordinate, not a rectangle, not a pixel rank.

Checked 2026-09-16 by reading each vendor's own product and pricing pages for the words pixel, rectangle and coordinate. The arithmetic here is only a count: twelve vendors examined, four publish a pixel-geometry field, eight do not. Naming a vendor in this table is a statement about what its pages published on that date and nothing more.

Why is a field this obviously useful this rare? Because ordinal position falls out of parsing a results page and a coordinate does not. A parser can tell you your result is the third organic link without knowing where the third organic link renders, how tall the block above it is, or whether something new appeared above it overnight. Coordinates only mean anything if the page has been laid out first and if every measurement uses the same viewport — and a consistent viewport is a product commitment, not a parsing detail. SerpApi says as much on its own page, describing pixel position as using an actual browser to determine element position on screen.

That is also why the feature is disproportionately valuable to you if you need it. In a category where twenty vendors compete on the same parsed fields, a field that eight of twelve do not sell is the one that decides the vendor, and price is what you argue about afterwards. If you are building the parsing side yourself, the companion walkthrough on parsing Google SERP features in Python shows exactly which blocks you would be locating by hand.

Engines, elements and devices

Among the four that do return geometry, coverage is where they separate. Three of them are Google-shaped; one covers four web-search engines. Read this before you assume a vendor's pixel feature follows its engine list — on two of the four it plainly does not.

Checked 2026-09-16 from each vendor's own feature page. The count that matters: of the four, one publishes the field for four web-search engines and three publish it for Google surfaces only. “Not stated for the feature” means the page I read named no engine list for the geometry — several of these vendors sell many engines, and the geometry may well follow them; their pages simply do not say so.

Devices are the row I could not fill honestly. None of the three competitor pages I read on 16 September 2026 stated a viewport, a device or a screen size for the pixel feature, so I am not claiming mobile support for any of them in either direction. What I can state is our side: Serpent's pixel fields are desktop-only today, measured against a 1366×768 viewport, and mobile is not available. Our Pixel Position API page carries a device-by-device comparison read on 7 September 2026 with more detail than I could verify in this pass — treat it as the fuller source, and re-check any device claim against the vendor before you build on it.

What the geometry costs

Now the part everyone scrolls to. Two of the four publish a dollar rate card you can read without talking to anybody, one publishes a rate card in pounds, and one — ours — bills the geometry at exactly the price of the search it rides on.

The arithmetic, checked 2026-09-16. DataForSEO prints its SERP prices per single task rather than per 1,000: $0.0006 for a queued Google organic SERP and $0.002 for the same task in Live mode. The DataForSEO help centre cost table charges the geometry parameter as base price plus base price, which puts a queued SERP with rectangles at $1.20 per 1,000 and a Live one at $4.00 per 1,000. SerpApi bills a pixel-position request as one search against the plan you are on, so its entry rate is $25.00 per 1,000 searches on the $25-a-month Starter plan, and the lowest rate its published ladder reaches is $1.96 per 1,000 searches on Cloud 54M at $106,050 a month. Serpent adds nothing for the flag: a web search with the pixel parameter costs the same $0.60 per 1,000 calls as one without, $0.06 on Growth and $0.03 on Scale. Authoritas is deliberately not converted, for the reason in the next paragraph.

The currency problem, and why there is no dollar column for Authoritas

Authoritas publishes its API pricing in pounds and nothing else. Its keyword ranking API page gives you 1,000 free API requests per month and then says pricing is volume-based starting at £3.50 CPM for 100,000 queries per month. CPM is cost per thousand, so the rate is £3.50 per 1,000 queries and the volume it starts at is 100,000 queries a month.

There is no dollar figure anywhere on that page, and I have not manufactured one. A conversion needs an exchange rate; an exchange rate needs a date; and a dollar figure published today against an undated rate is a number that will be quietly wrong within a month and will still be sitting in this table being quoted. So the pound stays a pound, in its own labelled column, and if you need the dollar equivalent you can apply today's rate yourself and know exactly which day it came from. That is less convenient and considerably more honest than the alternative.

Where these vendors beat us, plainly

Two of them beat us outright on getting started, and burying that would make this comparison worthless. Authoritas publishes 1,000 free API requests a month on the same API whose page advertises the (x,y) position of every element. SerpApi publishes a free plan of 250 searches a month, and publishes no surcharge anywhere on its pixel position page. A Serpent free-allowance call, by contrast, comes back with no pixel fields at all and a marker saying why, so a pixel job on our API has to be planned against paid credit from the first request. Neither of those vendors' pages states whether its free tier accepts the pixel output, so confirm it before you build a plan on it — but on the published evidence, they give you a way to try the geometry without paying and we do not.

DataForSEO also beats us on how little you have to commit before the rate applies: its published minimum payment is $50, and the queued price is the queued price whether you have spent fifty dollars or fifty thousand, because there is no volume ladder to climb. Our own lowest rate is a band, and reaching it means depositing more than that.

On price with the geometry switched on, the direction reverses. DataForSEO doubles the task for the parameter, which puts a queued Google SERP with rectangles at $1.20 per 1,000 SERPs and a Live one at $4.00 per 1,000 SERPs. SerpApi's entry plan works out at $25.00 per 1,000 searches, because the plan buys 1,000 searches for $25 a month and one pixel-position request spends one of them. Serpent bills the flag at the ordinary web rate, $0.60 per 1,000 calls by default and $0.03 per 1,000 calls once an account has reached the Scale band. The full ladder, and what each band actually requires, is on the pricing page.

So the geometry is not expensive anywhere. But on one of the four it is the difference between one base price and two, and that is worth knowing before you set the parameter on a million tasks and find out from the invoice.

Which option fits which job

• You already buy Google SERPs from SerpApi or DataForSEO. Use the parameter you already have. Adding a vendor to acquire one field you can already switch on is a procurement exercise, not an engineering win. On DataForSEO, budget for the doubling before you enable it at scale; on SerpApi, budget one search per pixel request against your existing plan.
• You need coordinates on more than Google. This is the decision the capability table makes for you: of the four, only the Serpent Pixel Position API publishes the field for Bing, Yahoo and DuckDuckGo as well as Google, and it costs the same on all four. If multi-engine geometry is the requirement, three of the four options are simply not candidates.
• You are doing share-of-voice or above-the-fold work. You need the height, so you need a rectangle. That rules out any option that publishes only an offset, and it is the reason the shape table above exists.
• You want to try it before paying anything. Authoritas at 1,000 free API requests a month, or SerpApi's 250-search free plan. Both beat our free allowance for this particular feature, which returns no pixel fields at all.
• You are building a rank tracker rather than a one-off audit. Store the coordinate next to the ordinal from day one; retrofitting it later means your history has a hole exactly where the interesting change happened. The rank tracking API and the argument in pixel position versus rank cover how the two series behave differently.
• You need a mobile viewport. Verify it with the vendor directly. None of the pages I read on 16 September 2026 stated one, ours included — and ours says plainly that mobile is not available today.
• You need the wider picture, not just this field. The full SERP API price comparison normalises the same vendors on ordinary web search, where the field list is common and price really is the whole argument. The Google SERP API page is the surface all four of these vendors agree on.

Whichever you pick, run the same two checks before you commit. Pull one real results page with the geometry on and confirm that the blocks you care about — not just the organic rows — actually carry coordinates. Then pull a results page with an AI Overview on it and confirm the same. Those two calls will tell you more than any comparison table, this one included.

FAQ

Which SERP APIs actually return pixel position in 2026?

Four, on the pages I read on 2026-09-16. SerpApi documents x, y, width and height on its Google Search API and its Google Ads API. DataForSEO returns a rectangle for each element when you set calculate_rectangles on a SERP task. Authoritas states on its keyword ranking API page that it can tell you the (x,y) position of all elements on the page. Serpent returns pixel_position and pixel_box on four web-search engines. Eight other vendors I checked — Bright Data, SearchApi, Oxylabs, HasData, Scrapingdog, Serper, Zenserp and Traject Data — publish no pixel-geometry field on their product pages, and that absence is the real headline of this comparison.

Does pixel position cost extra?

It depends entirely on the vendor, and it is the first thing to check. On DataForSEO the geometry is a paid parameter: setting calculate_rectangles adds one more base price to the task, so the same Google organic SERP goes from $0.60 to $1.20 per 1,000 on the queued method and from $2.00 to $4.00 per 1,000 in Live mode. SerpApi publishes no surcharge on its pixel position page, so a request with the pixel output is one search against whatever plan you are on. Serpent charges nothing extra either — a web search with pixel_position=true is billed exactly like one without, at $0.60 per 1,000 calls by default, $0.06 on Growth and $0.03 on Scale. Prices checked 2026-09-16.

How much does the calculate_rectangles parameter add to a DataForSEO task?

It doubles it. The DataForSEO help centre cost table, read on 2026-09-16, prices a queued Google organic SERP task at $0.60 per 1,000 and the same task with calculate_rectangles set to true at $1.20 per 1,000, because the parameter is charged as base price plus base price. In Live mode DataForSEO doubles the same way, from $2.00 per 1,000 SERPs to $4.00 per 1,000 SERPs. The high-priority row of that table came back malformed on both of my reads, so I am not quoting a priority-queue figure at all: that cell is empty in the table above and it should stay empty until somebody can read it cleanly.

Why do so few SERP APIs return pixel coordinates?

Because ordinal position falls out of parsing a results page and a coordinate does not. A parser can tell you that your result is the third organic link without knowing where the third organic link renders, how tall the block above it is, or whether something new was inserted above it since yesterday. Coordinates only mean anything if the page is laid out first and if every measurement uses the same viewport, which makes pixel geometry a product decision rather than a parsing detail. SerpApi says as much on its own page, describing pixel position as using an actual browser to determine element position on screen. That is why the field is rarer than the demand for it, and why eight of the twelve vendors I checked on 2026-09-16 publish nothing of the kind.

What does Authoritas charge for pixel position?

Authoritas publishes its API pricing in pounds only. Its keyword ranking API page gives you 1,000 free API requests per month and says that thereafter, pricing is volume-based starting at £3.50 CPM for 100,000 queries per month, where CPM means cost per thousand. There is no dollar figure on that page, and I have not converted it: a currency conversion with no dated exchange rate behind it is a number that rots quietly, and the rate is not mine to assert. The pound figure is the one that does not move. Checked 2026-09-16.

Which pixel position API should I choose?

Start from what you need rather than from price. If you only track Google and you already buy from SerpApi or DataForSEO, use the parameter you already have rather than adding a vendor. If you need coordinates on Bing, Yahoo or DuckDuckGo as well as Google, the Serpent Pixel Position API is the only one of the four that publishes the field on all four engines, at $0.60 per 1,000 calls and $0.03 per 1,000 calls at Scale, with no surcharge for the flag. If your volume is small and free matters more than breadth, Authoritas publishes 1,000 free API requests a month and SerpApi publishes a free plan of 250 searches a month, while a Serpent free-allowance call returns no pixel fields at all — though neither vendor's page says whether its free tier accepts the pixel output, so verify that before you plan around it. Checked 2026-09-16.

Related reading, in the order most people need it: start with what pixel position measures, then why rank and pixel position disagree, then how to score a SERP layout with the coordinates, and the API documentation when you are ready to send the first request.