Website Traffic Database — Visits, Rankings & Discovery

Find tracked websites that match a traffic range, niche, visitor country, channel mix, engagement or stored keyword sample. Look up known domains or discover category peers with comparable traffic.

The results are stored traffic estimates, not live measurements or the website owner's analytics. Check each row's latestMonth, dataDate and dataAgeDays before using it. A record checked recently can still contain an older metric month. Available categories, keywords and domains vary; no complete-web or fixed unique coverage count is guaranteed.

Start in three steps

1. Choose a workflow. Leave the traffic workflow selected for discovery.
2. Start with five websites and the Outreach output.
3. Read the Dataset and SUMMARY. Check the period, coverage and completion status before increasing the limit.

Small discovery input:

{
  "workflow": "traffic",
  "minVisits": 1000,
  "maxResults": 5,
  "outputPreset": "outreach"
}

Omitted maxResults now defaults to five for Console, API and MCP. Omitted outputPreset defaults to Outreach. Supply full explicitly if your integration needs nested keywords, countries or the monthly series.

Choose the workflow

Workflow	Selection	What it means
traffic	Visits range plus optional filters	Latest tracked monthly visits, not necessarily the current month
globalRank	topGlobalRank	Stored ranks 1–N; ascending rank; refine filters ignored
domains	domains	Only tracked domains; refine filters ignored
similarSites	similarTo	Same stored category, 25%–400% of seed monthly visits; seed excluded
auto	Existing input fields	Domains first, then similar seed, then top rank, then traffic

maxResults, sort, output and supported offset still apply to domain/peer discovery. Rank mode forces ascending global rank. Similar sites are category/traffic peers, not proven business competitors.

Copyable examples

Stored worldwide ranks — up to five rows:

{"workflow":"globalRank","topGlobalRank":5,"maxResults":5,"outputPreset":"compact"}

This selects ranks 1–5 if present in the stored data. It does not prove today's complete worldwide top five. Ties, missing ranks and source failures can change the row count.

Look up three tracked domains:

{"workflow":"domains","domains":["https://www.google.com/","youtube.com","facebook.com"],"maxResults":3,"outputPreset":"full"}

URLs normalize to hostnames, including removal of www; duplicates are removed. Credentials, ports, IP addresses and local hosts are rejected. Untracked domains produce no invented row. An empty result says nothing about the site's real traffic.

Finance shortlist with US as the top visitor country:

{"workflow":"traffic","category":"finance","country":"US","minVisits":100000,"maxResults":5,"outputPreset":"outreach"}

Country is the stored leading visitor market, not the company's address.

Growing sites with search and AI referral traffic:

{"workflow":"traffic","minVisits":100000,"onlyGrowing":true,"minSearchShare":10,"hasAiTraffic":true,"maxResults":5,"outputPreset":"full"}

Growth compares the earliest and latest tracked month. It is not necessarily a one-month change. Unknown or zero starting visits do not qualify. AI share and aiVisits are distinct stored metrics; do not calculate one from the other without a compatible period.

Find category peers:

{"workflow":"similarSites","similarTo":"bing.com","maxResults":3,"outputPreset":"compact"}

This requires the seed's stored category and positive monthly visits. Missing seed metadata is explained in ERRORS.

Restrict by record check date:

{"workflow":"traffic","checkedWithinDays":30,"maxResults":5,"outputPreset":"compact"}

Older records and future check dates are excluded. A valid empty result is possible. This does not refresh the underlying traffic estimate.

keyword matches a literal substring in the available top-keyword names; it is not an exhaustive ranking index or live SERP query. Combine it with category or traffic constraints. % and _ are treated as text, not wildcard commands.

Sorting and continuation

For limits up to 50,000, the Actor merges accessible source candidates in global order, with domain as a stable tie-breaker and unknown sort values last. Filtering, including positive growth, happens before source limits. Duplicates are removed before billing.

When SUMMARY.nextInput is present, copy that object into your next run. It advances from the number of rows actually delivered, including when a charge limit stops a run. Do not infer an offset from the number requested.

offset is capped at 5,000; offset plus maxResults must stay within 50,000. Data can change between runs; no snapshot is retained. Continuation is suppressed after source failures or invalid source rows. A returned next input is a way to continue, not a promise that the next page contains rows.

For limits above 50,000, bulk export streams in segment order. It is sorted within each segment and is not globally ranked. SUMMARY.globalOrder is false and sortScope is within_each_segment. The ceiling is 1,000,000 delivered rows, subject to available data, time and charge limits. Bulk mode does not support offset continuation. A small run does not verify a million-row export; use a deliberate budget for large pulls.

Results and evidence

Field	Meaning
domain	Normalized tracked hostname
latestMonthVisits, latestMonth	Stored visits and their month
totalVisits, monthlyVisits	Stored tracked-period total and series, where available
monthlyGrowthPercent, growthDefinition	Earliest-to-latest tracked-month change
globalRank, categoryRank, countryRank	Stored ranks; unknown remains null
category, country	Stored category and top visitor country
bounceRate, pagesPerVisit, avgTimeOnSite	Estimated engagement; time in seconds
trafficSources	Stored percentage shares, including AI where available
aiVisits	Stored AI referral visit estimate; not derived from channel share
topKeywords, topCountries	Available samples/breakdowns in Full output
dataDate, dataAgeDays, dataDateStatus	Record check date, age and date validity
dataSource, isLiveMeasurement	Stored estimates; live measurement is false

Missing numeric values remain null; genuine zero remains zero. Unknown dates have no fabricated age; future dates are flagged. Full output preserves the source's nested series and samples; use JSON for those fields. Dataset exports support CSV and Excel; flatten nested fields for spreadsheet analysis.

Compact retains core metrics and evidence. Outreach adds engagement/channel mix. Full adds stored monthly history, top countries and top keywords.

Check completion

Every run writes SUMMARY and ERRORS; successful collection also writes CONTINUATION. Diagnostic records are separate from billed website rows.

• SUCCEEDED: useful rows delivered without detected source/quality/budget failures; a result limit can still truncate the query.
• NO_RESULTS: accessible queries returned no useful matches. It does not establish zero real traffic or complete-web absence.
• PARTIAL: useful data may remain, but a source failure, skipped invalid row or charge limit makes the run incomplete.
• FAILED / INVALID_INPUT: collection or validation failed; inspect diagnostics before retrying.

Inspect sourceCoverage, coverageComplete, queryExhausted, chargeLimitReached, delivered and nextInput. Source failures are sanitized and never represented as healthy empty segments. Coverage means the configured sources were queried; it does not prove coverage of every website. Bulk runs stopped at a limit can leave segments unqueried.

MCP and AI agents

Use this tool to discover traffic-qualified tracked websites or retrieve their stored metrics. It is unsuitable for a claim about today's live traffic.

Select this Actor in API → MCP, or connect an authenticated client to:

https://mcp.apify.com/?tools=apivault_labs/website-traffic-database

Direct tool: apivault_labs--website-traffic-database. Keep the Apify token in secure client configuration. No external traffic-service key is required from the user; Apify authentication and run charges still apply.

Read tools/list for the current wrapper. Send a copyable input above as the direct tool arguments; the wrapper can additionally accept waitSecs: 45. That wait is not the run timeout. If the run is still active, poll the returned run ID with get-actor-run; do not launch a duplicate.

Once terminal, read get-dataset-items using the actual returned dataset ID, a small limit and fields such as domain,latestMonthVisits,latestMonth,dataDate,dataAgeDays. Read SUMMARY and ERRORS with get-key-value-store-record using the actual returned store ID. Missing storage IDs are an error to report, not a reason to use another run's data.

Inspect status, coverage, period and missing values before answering. Use SUMMARY.nextInput for authorized continuation. Runtime/memory/charge controls belong in Console or a run client that actually exposes them; they are not arbitrary Actor input fields.

Agent request: “Find five finance websites with at least 100,000 stored monthly visits and US as their top visitor country. Return domains, metric month and record age; explain incomplete source coverage or older data.”

Pricing

$0.005 per delivered website ($5 per 1,000), on all current plans. There is no Actor start event. Apify compute, storage and other platform usage are additionally paid by the user. Check current Pricing before execution.

Only useful rows accepted for delivery are billed as result events. Source errors, missing domains and diagnostic records are not billed as website results. A zero-row run can still incur platform usage. Set a small maxResults and a run charge cap; large limits do not imply permission to spend more.

FAQ

Why is a familiar site missing? It may not be tracked, may lack required metrics, may not match your filters, or may be in a failed source segment. Read SUMMARY and ERRORS.

Does running it update the website's traffic? No. generatedAt is the run time; dataDate is the stored record check date; latestMonth identifies the stored metric period.

Are all competitors listed? No. Similar-site mode is a category/traffic heuristic over available stored data.

Does a global-rank query guarantee current worldwide rankings? No. It selects stored ranks from accessible data and reports coverage.

Can I use this without logging into another traffic service? Yes. The Actor handles its data access; you still need Apify access and authentication for API/MCP use.