How complete, fast and expensive are SEO data APIs on real workloads?
Measured comparisons of SEO API output on fixed query cohorts. One declared cohort, every route on the same day, every raw response published as data you can download and re-run.
50
Queries in cohort
20
Domains
8
Routes measured
314
API calls · 284 valid
$1.99
Run cost, list price
1
Run · latest 2026-09-15
Catalogue
| Record | Type | Cohort | Run | Calls | Files |
|---|---|---|---|---|---|
| Pay per lookup for occasional SEO audits: what one client audit cost through DataForSEO and Serpstat, measuredOne client audit (rankings for 50 queries, volumes for the same 50, the domain's ranked keywords, its backlink profile and a 100-page on-page crawl) run through two pay-per-call routes on 29 October 2026, with the settled cost per step and what the API did not return that a suite's screens do. | Field guide | v1 | 2026-09-15 | 314 | csv · raw |
| Does SERP keyword clustering depend on the SERP API? 50 keywords, six runs, measuredThe rule people give Claude Code with a SERP API: group keywords whose top 10 results share 5 or more URLs. Applied to the same 50 keywords through nine SERP routes across six runs between 15 September and 15 October 2026, how many keyword pairs each route groups, how far the groups move between runs half an hour apart, and what 1,000 keywords cost per route. | Field guide | v1 | 2026-09-15 | 314 | csv · raw |
| How to check an AI visibility tracker is pulling live answers: seven answer routes, measuredTen prompts, four of them about September 2026 events, through ChatGPT, Perplexity, Google AI Mode and Gemini, and through the ChatGPT model API with web search off and on. Which answers were live, which came from model memory, and what a tracker has to store to prove the difference. | Field guide | v1 | 2026-09-15 | 314 | csv · raw |
| Are SERP APIs blocked by Google's scraper crackdown? Nine routes, measuredThe same 50 queries through Google SERP routes on 15 and 28 September 2026: how many calls failed, whether the google.com/goto redirect reaches the URL field, and why routes that agreed on 15 September now return different result lists for the same query. | Field guide | v1 | 2026-09-15 | 314 | csv · raw |
| What SEO data you can actually get for free, and where free runs outFree tiers, one-off grants and trials for Google results, keyword volume and backlinks, read from each provider's own pricing page on one day: how many calls, for how long, and the first paid step. | Field guide | v1 | 2026-09-15 | 314 | csv · raw |
| Is there an API for Google AI Overviews? What three routes actually returnGoogle publishes no AI Overview API. Three third-party routes claim to return the block: a rendered-page capture, a SERP API flag, and a two-step token exchange. Measured on 90 and 100 requests on the same day: how often the overview came back, with what, and at what price. | Field guide | v1 | 2026-09-15 | 314 | csv · raw |
| Connect Search Console and keyword research to Claude CodeHow a coding agent reads a site's Search Console data and pulls keyword volumes, SERPs and backlinks through one connection instead of four integrations: the actual calls, what they returned, and what they cost. | Workflow | v1 | 2026-09-15 | 314 | csv · raw |
| Backlink counts across providers: same ranking, different numbersTwenty vendor domains queried on Serpstat, SE Ranking and Moz on the same day. How far the referring-domain counts diverge, whether the providers at least agree on order, and why authority scores cannot be compared across tools. | Benchmark | v1 | 2026-09-15 | 314 | csv · raw |
| Keyword volume coverage: which APIs report the long tail at allFifty US-English buyer keywords sent to DataForSEO, Serpstat, SE Ranking and Google Ads Keyword Planner in one call each. How many came back with a number, how the numbers compare with Google's own, and how each provider signals 'no data'. | Benchmark | v1 | 2026-09-15 | 314 | csv · raw |
| SEO API latency, failures and what a real workload costsLatency, failure and no-data rates, and observed cost for 314 calls across twelve SEO data routes on one day, with a worked monthly workload for an in-house growth agent. | Benchmark | v1 | 2026-09-15 | 314 | csv · raw |
| Google SERP completeness: live fetch versus stored indexFifty queries fetched through five Google results routes on the same day and compared with a live rendered Google page: how much of the real top 10 each route returns, whether the first result matches, what it costs and how long it takes. | Benchmark | v1 | 2026-09-15 | 314 | csv · raw |
| SEO API evaluation cohort v1: reproduce the benchmarkThe declared cohort of 50 queries and 20 domains, the routes it was sent through, the exact commands, and the per-call record format, so anyone can repeat the run and diff it against ours. | Dataset | v1 | 2026-09-15 | 314 | csv · raw |
Latest run
2026-09-15 · cohort v1 · 314 calls, 284 valid, $1.99.
Routes: dataforseo, google-ads, seranking, serpstat, cloro, serpapi, treg-routed, moz.
Not measured: Ahrefs (plan), Majestic, Semrush (no credential).
How to cite
Lou, S. SEO Data Bench: SEO API evaluation cohort v1, run 2026-09-15. https://seodatabench.com/
Provenance
Every number links to a dated raw response. Records that include a route the publisher has a relationship with say so on the page. Methodology · Relationships