Microsoft Clarity MCP Server & Skill

Tools: 5

Microsoft Clarity

Connect your account, then chat with AI to run tools.

Query Microsoft Clarity traffic and friction metrics, ask dashboard questions in plain language, list session recordings, and search official docs — all with one project token.

Microsoft Clarity is used here only to identify this integration; MCPBundles is not affiliated with or endorsed by Microsoft Clarity or its owner.

Best forGrowth

Live probe refreshed Sep 18, 2026 · Endpoint host mcp.mcpbundles.com

What you can do with Microsoft Clarity

Built for

Growth Teams, Product Analytics, Conversion Optimization, Marketing Operations

Example workflows

Yesterday's traffic snapshot

Structured export for a daily pulse — counts against the 10/day export cap.

Using Microsoft Clarity, export the last 24 hours of session traffic and engagement time for our site and summarize whether bot traffic is material.

Setup Instructions

Connect Microsoft Clarity to any MCP client in minutes

Use in your AI tool

You’ll sign in to MCPBundles when your client connects.

MCP URL

https://mcp.mcpbundles.com/bundle/microsoft-clarity

Claude.ai Cursor VS Code VS Code Insiders Goose LM Studio Cherry StudioChatGPT

Chat in MCPBundles Studio

This server needs a sign-in. MCPBundles handles that setup once — then chat here with saved access and a log of what ran.

Ready to use Microsoft Clarity?

Sign in to connect your credentials and start running tools from the chat.

MCPBundles publishes this directory for MCP discovery. Except where we host or operate an offering, third-party services run under their own terms. Product and company names on this page are used in a descriptive, identifying way (including under nominative fair use where applicable); they remain the property of their owners. Nothing here grants you rights in those marks, and nothing here is an offer to sell a third party's services. Terms

AI Skill
SKILL.md

Domain knowledge for Microsoft Clarity — workflow patterns, data models, and gotchas for your AI agent.

Microsoft Clarity

Microsoft Clarity records how visitors use a website: sessions, scroll depth, engagement time, rage clicks, dead clicks, quick backs, script errors, and traffic sources. One Data Export token (Clarity → Settings → Data Export) powers all tools in this bundle.

Four API surfaces (same token)

Surface	Backend	Best for
Structured metric export	clarity.ms/export-data GET	Deterministic metric blocks with dimension_1–dimension_3 and metric_names filtering
NL dashboard query	clarity.microsoft.com/mcp/dashboard/query POST	Natural-language dashboard questions ("top browsers last 3 days")
Session recordings sample	clarity.microsoft.com/mcp/recordings/sample POST	Sample session recording metadata (links, duration, filters)
Documentation search	clarity.microsoft.com/mcp/documentation/query POST	Official Clarity documentation snippets
Snapshot inventory	(local inventory)	See cached upstream payloads and quota without calling Microsoft

Snapshot caching (all fetch tools)

Every upstream tool accepts refresh and allow_stale:

Default: serve a fresh cached snapshot when one exists for the same upstream arguments (dimensions, date range, NL query, etc.).
metric_names and compact on structured export are local filters — they narrow a cached export without spending quota.
When export quota is exhausted, set allow_stale=true (default) to read the latest cached export; only use refresh=true when the user explicitly wants a new upstream call.
Snapshot inventory lists stored snapshots and quota windows for the connected credential.

Operational order

Snapshot inventory or a cached read when you only need data already fetched for this credential.
Documentation search when the user needs setup, SDK, privacy, or feature guidance — no export quota cost.
Structured export with num_of_days=1 and no dimensions for a traffic pulse — only 10 calls/project/UTC day on this endpoint.
NL dashboard query for exploratory questions when structured export is awkward — separate from the export quota.
Session recordings sample to find sessions with rage clicks, errors, or device filters — empty [] is normal on low-traffic projects.

For structured breakdowns, prefer one export call with dimension_1 set rather than multiple aggregates. Use metric_names to trim a response locally without another export call.

Metrics from structured export

Metric	Meaning
Traffic	Session counts, bot sessions, distinct users, pages per session
EngagementTime	Total and active engagement time
ScrollDepth	Average scroll depth
DeadClickCount	Clicks on non-interactive elements
RageClickCount	Rapid repeated clicks signaling frustration
QuickbackClick	Fast navigations away after landing
ExcessiveScroll	Unusually long scroll sessions
ScriptErrorCount	JavaScript errors observed in sessions
ErrorClickCount	Clicks that coincided with errors
Browser / Device / OS / Country	Dimensional breakdown rows when requested
PageTitle / ReferrerUrl / PopularPages	Content and acquisition breakdown rows

Quotas and limits

Structured export only: 10 requests per project per UTC day (HTTP 429 Exceeded daily limit). History window is 1–3 days; max 1,000 rows; no pagination; timestamps are UTC.
NL dashboard, recordings, and docs: use Microsoft's clarity.microsoft.com/mcp HTTP APIs with the same token. They are not blocked when the export daily cap is hit — plan agents around that split.
Session recordings return metadata and links, not video files. Heatmap images are not exposed via API.
Clarity is free; there is no paid tier to raise export limits today (Microsoft has noted higher limits on their roadmap).

Gotchas

Low-traffic projects return HTTP 200 with zero/null metrics or empty recording arrays — use compact=true (default) on exports to drop empty sections.
Each token is scoped to one Clarity project; agencies need one credential per project.
NL dashboard queries work best with single-purpose questions and explicit time ranges; avoid multi-metric kitchen-sink prompts.
Invalid dimension or metric names are rejected before the upstream export call.