Read-only MCP Server for Web Analytics

HitKeep is a read-only MCP server for web analytics when an assistant needs aggregate traffic, events, goals, ecommerce, Search Console, Web Vitals, and AI crawler visibility without dashboard cookies or raw hit exports. Connect an MCP client to HitKeep’s configured HTTPS MCP endpoint with a scoped API-client bearer token, or use OAuth where the instance advertises support. Check release availability and instance support before using OAuth or the Assistants setup page shown below.

Use this page when you are evaluating an analytics MCP server, adding HitKeep to an MCP directory, or deciding whether an assistant should query analytics through MCP instead of the dashboard or REST API.

HitKeep Assistants integration page showing the MCP endpoint and client setup guidance — Copy connection details from Integration → Assistants, then authorize a compatible MCP client with scoped access to aggregate analytics.

Best Fit

Use this path when you want:

Claude, Cursor, or another MCP client to answer analytics questions from approved HitKeep data
internal reporting tools that can query aggregate site metrics without dashboard access
scoped assistant access by site through OAuth consent or API-client tokens
traffic, event, ecommerce, Web Vitals, AI visibility, Search Console, Opportunities, and docs lookup through one governed interface
a clear boundary that keeps raw exports, write workflows, admin actions, and dashboard sessions outside MCP

If your assistant supports Agent Skills, pair MCP with the official HitKeep Agent Skills. The skills teach the assistant which MCP tools to use for traffic diagnosis, AI visibility, ecommerce analysis, Search Console checks, and tracking verification.

Quick Connect

For self-hosted HitKeep, enable MCP on the leader:

hitkeep \
  -mcp-enabled \
  -mcp-path /mcp

Equivalent environment variables:

HITKEEP_MCP_ENABLED=true
HITKEEP_MCP_PATH=/mcp

If HitKeep is served at https://analytics.example.com, the MCP endpoint is:

https://analytics.example.com/mcp

If the instance advertises OAuth and your client supports it, connect to the canonical HTTPS MCP URL and approve sites:read analytics:read for one team and selected sites. Review your grant by opening the header user menu and choosing Connected applications. A team owner or admin can register public PKCE clients at Integration → OAuth clients when the client requires a registered ID. See OAuth setup.

For a client without OAuth, send a manually created API-client token as a bearer header:

Authorization: Bearer <hitkeep-api-client-token>

MCP client configuration formats differ. A manual bearer-token connection has this shape:

{
  "mcpServers": {
    "hitkeep": {
      "type": "streamable-http",
      "url": "https://analytics.example.com/mcp",
      "headers": {
        "Authorization": "Bearer ${HITKEEP_MCP_TOKEN}"
      }
    }
  }
}

For reverse proxy details, leader-only behavior, host validation, date-range limits, and audit commands, use the official MCP server guide.

Create A Scoped Token

Use a personal or team API client token rather than a dashboard session:

Open Integration → API for a personal token, or Administration → Team → API clients for a team token.
Grant only the sites the MCP client should query.
Copy the token into the MCP client’s secret store.
Rotate or revoke the token when the assistant no longer needs access.

HitKeep API clients page showing scoped bearer tokens for programmatic access — API clients provide manual bearer tokens for MCP clients without OAuth. Dashboard cookies are not accepted by the MCP route.

Team API clients are the better default for shared assistants because the token is owned by the team and can be revoked without touching a person’s dashboard account.

Exact MCP Tools

HitKeep MCP exposes read-only tools for aggregate analytics, imported Search Console rows, official docs lookup, and local MCP help.

Tool	What it returns
`hitkeep_list_sites`	Sites authorized for the OAuth grant or API client.
`hitkeep_get_site_overview`	KPI totals and optional period comparison; sole `sections:["kpis"]` omits optional charts, lists, and goals.
`hitkeep_get_breakdown`	Pageviews and visitors by one dimension with optional filters; comparison shares use the fetched top 500 tracked values per period, with explicit coverage metadata.
`hitkeep_get_funnel_stats`	Step visitors, drop-off, and conversion for a configured funnel; obtain its `funnel_id` from the overview’s `goals` section.
`hitkeep_get_qr_campaigns`	A site’s QR campaign opens, pageviews, visitors, and optional open timeseries, without redirect tokens or raw hits.
`hitkeep_get_event_names`	Event names for a site and date range.
`hitkeep_get_event_breakdown`	Ordered property-key discovery when `property_key` is omitted; capped value breakdowns when a nonblank key is supplied.
`hitkeep_get_ecommerce`	Ecommerce summary, products, source stats, city, provider, and ASN aggregates, and an optional timeseries.
`hitkeep_get_web_vitals`	Aggregate Web Vitals p75 values, sample counts, rating counts, optional page rows, and optional visitor-context breakdowns.
`hitkeep_get_ai_visibility`	AI crawler fetch overview and timeseries, with optional correlation context.
`hitkeep_get_annotations`	The team’s chart annotations for a date range, such as releases, campaigns, and outages, without author information.
`hitkeep_get_opportunities`	Saved Opportunity recommendation records for a visible site.
`hitkeep_get_search_console_status`	Search Console mapping, sync status, imported date range, data availability, and safe attention reason.
`hitkeep_get_search_console`	Imported Search Console overview and series, plus requested query, page, country, or device dimensions.
`hitkeep_search_docs`	Search the cached HitKeep docs catalog and fetched markdown snippets.
`hitkeep_get_doc`	Fetch one official HitKeep docs page as markdown.
`hitkeep_get_api_reference`	Fetch REST API reference markdown for a path or operation slug.
`hitkeep_get_mcp_help`	Return local MCP setup, date range, filter, and privacy guidance.

Resources exposed to clients:

Resource	What it is for
`hitkeep://help/mcp`	Token, date range, filter, and privacy guidance.
`hitkeep://help/metrics`	Definitions for core metrics returned by MCP tools.
`hitkeep://docs/llms`	Cached official docs catalog.
`hitkeep://docs/{path}`	Markdown for a specific official docs page.

Permission Boundary

HitKeep MCP is designed for assistant reads, not assistant administration.

Boundary	HitKeep behavior
Default state	MCP is disabled until enabled by the operator.
Transport	Streamable HTTP on the existing HitKeep server, usually `/mcp`.
Auth	API-client bearer tokens, or resource-bound OAuth access tokens on instances that advertise support.
Cookies	Dashboard cookies are not accepted.
Site access	The token must have a site grant and pass normal `site.view` checks.
Data access	Aggregate analytics, imported aggregate Search Console rows, saved Opportunities, and docs tools.
Excluded	Raw hit exports, raw visitor rows, dashboard sessions, billing, token management, site administration, goal mutation, funnel mutation, exclusions, and alert creation.

Analytics tools default to the last 30 days when no date range is supplied. Date ranges and result limits are capped by the MCP configuration. Filters support common aggregate dimensions such as path, hostname, referrer, device, country, city, provider, ASN, browser, language, and UTM fields.

Use open exports and takeout when a person needs portable files or raw records that are available through export workflows. MCP intentionally stays narrower.

Google Search Console Analytics Through MCP

HitKeep imports finalized, aggregate Search Analytics rows from Google Search Console. MCP can then expose those imported rows to approved assistants through hitkeep_get_search_console_status and hitkeep_get_search_console.

HitKeep Search Console dashboard with clicks, impressions, CTR, average position, queries, pages, countries, and devices — Search Console MCP tools read the same imported aggregate rows that power the authenticated dashboard drilldown.

The Search Console MCP tools do not call Google live, refresh OAuth credentials, or trigger a sync. Empty reports should be checked with hitkeep_get_search_console_status before drawing conclusions, because a site can have HitKeep traffic while Search Console import is not mapped or not yet synced.

For setup and troubleshooting, read the Google Search Console integration guide.

AI Crawler Analytics Through MCP

Use hitkeep_get_ai_visibility when an assistant needs aggregate AI crawler analytics: fetch volume, assistant families, top paths, failures, AI-referred visits, and optional fetch-to-visit correlation.

HitKeep AI visibility dashboard with AI crawler fetch KPI cards, assistant filters, and crawl demand reporting — AI visibility keeps AI crawler fetches and AI-referred human visits separate so assistants do not treat crawl volume as demand.

Browser tracking can capture many AI-referred human visits. AI crawler fetches usually require a server-side feed from logs, middleware, a CDN, or an edge worker because crawlers often do not run browser JavaScript.

HitKeep can show crawler interest and later AI-referred visits on the same paths. It cannot prove a specific answer, citation, or ranking unless that evidence exists in the data source.

When To Use Another Surface

Need	Better surface
Human setup or visual investigation	Dashboard
Normal product automation	REST API
Raw files, audit handoff, or account export	Open exports and takeout
Assistant analysis over approved aggregate reports	MCP plus HitKeep Agent Skills
Managed endpoint, backups, updates, and email delivery	HitKeep Cloud

FAQ

Is HitKeep MCP read-only?

Yes. HitKeep MCP exposes aggregate analytics and docs tools only. It does not expose write workflows, admin actions, billing actions, token management, dashboard sessions, or raw hit exports.

Can HitKeep MCP read Google Search Console analytics?

Yes. HitKeep MCP can read imported aggregate Search Console rows for a site. It does not call Google live, refresh OAuth credentials, or trigger a sync.

Can HitKeep MCP analyze AI crawler traffic?

Yes. The AI visibility tool can return aggregate AI crawler fetches, AI-referred visits, and optional correlation context when the site has AI fetch data.

Does HitKeep MCP accept dashboard cookies?

No. Use API-client bearer tokens with explicit site grants, or scoped access tokens for the MCP resource on an OAuth-enabled instance. Keep credentials in the MCP client’s secret store.