Similarweb API Pricing and Cost: Quote-Only Credits
Similarweb API pricing is quote only. It sells credits per data point, Business or standalone, with a public formula to size your bill before the sales call.
By the AdEye team
September 2026 · 8 min read
Searching the Google Ads Transparency Center...
Public Google Ads Transparency data. Nothing behind a login.
Showing of live ads from Google Ads Transparency live ads from Google Ads Transparency
Similarweb does not publish an API price. The API is sold inside a custom Business or Enterprise package, or as a standalone API package without a platform subscription, and either way the price comes from sales. What Similarweb does publish is the meter: you pay in data credits for the data points returned, not for the calls you make, and unused credits expire at the end of each billing cycle. Checked on similarweb.com and docs.similarweb.com in September 2026.
So a "Similarweb API pricing" figure you read on a roundup site is either a guess or a price for something else. What you can do before the sales call is work out how many credits your project needs, because the formula is public and precise. That number is the thing you negotiate on. Below is how the credit math works, what the API actually returns about competitor advertising, the limits that shape a build, and how it compares with the other ad data APIs a US team would price alongside it.
How much does the Similarweb API cost?
There is no list price. Similarweb's own pricing FAQ says the API is available as part of its customized packages for businesses, or as a standalone product, and that to learn about pricing you contact sales and select "API" in the form. None of the three self-service Web Intelligence plans ($199, $399 and $649 a month, or $125, $333 and $542 a month billed annually) lists API access. Data integrations, real-time and bulk APIs, BI connectors and MCP sit in the Businesses and Enterprises column instead.
| How you buy it | Published price | What it includes | Who it suits |
|---|---|---|---|
| Self-service Web Intelligence plans | $199, $399 or $649 a month | The web app, one seat. No API access listed | Analysts who work in the interface |
| Business or Enterprise package with API | Custom quote | Platform seats plus a monthly data credit allowance | Teams that want the app and a pipeline |
| Standalone API package | Custom quote | Data credits and an API key, no platform subscription | Developers feeding a warehouse or product |
The practical consequence: there is no self-service route to the API, so plan the sales conversation around a credit volume rather than a plan name. The self-service plans themselves are broken down tier by tier in our Similarweb pricing guide.
How do Similarweb API data credits work?
Every data point an endpoint returns costs a set number of credits. Similarweb's documentation gives the formula as: Total credits = domains x endpoint base cost x granularity multiplier x country filters x time range x results per domain. Standard endpoints, such as Traffic and Engagement, Traffic Sources and Demographics, cost 1 credit per data point. Keyword endpoints, including Paid Keywords and Landing Pages, cost 1 credit per 100 data points.
Granularity multiplies the bill quietly. Monthly data returns one result per month, weekly about four, daily about thirty. The same question asked daily instead of monthly costs roughly thirty times as much, which is an easy way to burn a month of credits in the first week.
Similarweb's own two examples show the spread. Six months of monthly US visits for amazon.com costs 6 credits. Three months of weekly traffic sources for walmart.com across the US and Canada costs 576 credits, because traffic sources returns 8 results per period and the request doubles for the second country.
Three rules protect you. Error responses (401, 403, 429 and other 4xx or 5xx errors) are not charged, and neither is a request that finds no data. There is a Request Validate endpoint that estimates the credit cost of a query before you run it. And there are endpoints to check remaining credits for the account and for each user, so a script can stop before it overdraws.
What does a competitor ad research project cost in credits?
Here is a realistic US setup: 20 competitors, one country, refreshed monthly. These are credit counts from the published formula, not dollars, because Similarweb does not publish a dollar rate per credit.
| Pull | Math | Credits per month |
|---|---|---|
| Monthly visits and engagement, 20 domains | 20 x 1 x 1 month x 1 country | 20 |
| Traffic sources split, 20 domains, monthly | 20 x 8 results x 1 month | 160 |
| Same traffic sources pulled weekly | 20 x 8 results x about 4 weeks | about 640 |
| Top 5,000 paid keywords per domain | 20 x 5,000 data points / 100 | 1,000 |
| Paid landing pages, 500 per domain | 20 x 500 / 100 | 100 |
A monthly competitor dashboard of that shape lands around 1,300 credits, and around 1,800 if you insist on weekly traffic sources. A one-off backfill of 12 months multiplies the time-based rows by 12. Walk into the sales call with a table like this, and ask for the credit allowance and the price per extra block of credits in writing. Because credits expire every cycle, buying far above your real run rate is money left on the table.
Does the Similarweb API include competitor ad data?
It includes advertising metrics, not ads. The v5 endpoint reference lists PPC Spend (an estimate of a site's monthly paid search spend, in USD, EUR, GBP, AUD or JPY), Paid Keywords, Ad Networks, Publishers and Outgoing Ads, and the Landing Pages endpoint carries an "ads" field when you filter to paid traffic. That field is documented as the number of unique ad creatives or variations observed for a URL. It is a count. It does not return the headline, the image or the video.
Two things follow for a buyer. First, spend figures are modelled. Nobody outside an advertiser's own account reads their real Google Ads spend, Similarweb included, so treat PPC Spend as a directional estimate you compare across rivals rather than a number you put in a board deck. Second, if the brief is "show me what our competitors are running," the Similarweb API answers who, where and roughly how much, and leaves the creative question open.
What are the Similarweb API rate limits and history limits?
| REST API | Batch API | |
|---|---|---|
| Request style | Synchronous GET, answers in seconds | Asynchronous job, minutes to hours |
| Format | JSON | CSV |
| Domains per request | Typically 1 to 100, by endpoint | Up to 1,000,000 |
| History | Up to 37 months | Up to 7 years |
| Rate limit | 10 requests per second | No limit on job submission |
| Warehouse delivery | No | S3, Snowflake, Databricks |
Both run on the same API key. The history you can reach also depends on your package, and a capabilities endpoint tells you what yours covers. When you go past the monthly allowance, requests fail with a 429 rather than billing you on, which is safer than an uncapped card but means a scheduled job can silently stop mid-month. If Batch files land in your warehouse, it pays to have data lineage that flags when a scheduled feed goes stale, so a missed refresh shows up before someone builds a report on last month's numbers.
Is there a free Similarweb API?
Not a general one. Similarweb offers a free trial on its self-service plans, but those plans do not list API access, and API access is arranged through sales. Listings for a "Similarweb API" on third-party API marketplaces with $19 or $49 plans are published by independent accounts, not by Similarweb, so check who you are paying and what the data source is before you wire one into a product.
Similarweb API vs other ad data APIs
| API | Published price | Ad data returned | Main limit |
|---|---|---|---|
| Similarweb | Quote only, credit based | Modelled PPC spend, paid keywords, ad networks, ad counts per landing page | No dollar rate per credit published. No creative returned |
| Semrush | $549 a month Advanced plan plus API units, unit price unpublished | Google paid keywords, text and Shopping ad copy | Modelled data, text copy rather than live creative |
| SpyFu | $0.20 to $5.00 per 1,000 rows, on the $119 and $249 plans | Google keywords, competitors, ad copy text history | Google search only, modelled |
| Google Ads Transparency Center (official) | Not sold commercially | Ads served in the EEA | No endpoint a US marketer can buy |
Of the three commercial options, SpyFu is the only one where you can price a project on a whiteboard, because its rate card is public. Semrush and Similarweb both require a conversation. Our Semrush API pricing breakdown and SpyFu API pricing guide carry the unit tables for each, and the ad library API overview covers what the official platform APIs do and do not expose.
Is the Similarweb API worth it for competitor ad research?
It is worth it when your question is market level: which rivals are growing, how their traffic splits between paid and organic, which ad networks and publishers they lean on, and roughly how their search spend moves month to month, across dozens or thousands of domains. Batch delivery into Snowflake or S3 with seven years of history is a genuine strength that the ad spy tools do not match.
It is the wrong purchase when the job is creative. A credit contract that returns "14 ad variations observed" for a landing page does not show your team a single one of them. For that half, the source is Google's own Ads Transparency Center, which shows the ads a brand is running across Search, Shopping, Display and YouTube. AdEye reads it for you: search a competitor by name or domain with the tool above and you get their live Google ads, tagged by hook and angle and ranked by how long each has run, on plans from $39 a month, or $19.50 a month billed yearly. What it covers, and what it does not, is set out on our Google Ads spy tool page.
The two answer different questions, and plenty of teams run both: Similarweb to see where a competitor spends, AdEye to see what they say when they get there.
See what your competitors are running with AdEye
Type any brand and watch their live Google ads stream into one feed, each AI-tagged with its hook, angle and a working signal. Public ad-library data only.
Popular ad spy use cases