Measure MCP Server: Setup for Claude, Cursor and Other Apps

• Asks Measure Agent a free-form question about your app data

  Hands a plain-language question to Measure's own agent, which picks the right tool or runs read-only queries on raw telemetry and answers with numbers. Each call stands alone, so a follow-up should repeat the earlier findings.

  Try: “Which release has the most new crashes this week, and what share of sessions does it hit?”

  Read · ask_question

• Lists the apps you can see, with their team

  Returns every app your account can reach and the team each belongs to. The assistant needs these app IDs for almost every other call.

  Try: “Which apps do I have in Measure?”

  Read · list_apps

• Lists the fields you can filter on

  Returns the filter vocabulary for one data type such as errors, sessions or spans, with each field's type and allowed operators. The assistant calls it before writing a filter.

  Try: “What can I filter crashes by?”

  Read · get_filter_keys

• Lists values a filter field can take

  Returns suggested values for one filter field in one app, such as the app versions or countries it has seen.

  Try: “Which app versions have reported errors?”

  Read · get_filter_values

• Gets crash-free rates and launch times

  Returns adoption, crash-free, ANR-free and app-hang-free session rates and cold, warm and hot launch p95 for an app.

  Try: “What is the crash-free session rate for our Android app?”

  Read · get_metrics

• Shows sessions, crashes, ANRs and hangs over time

  Returns counts of sessions, fatal exceptions, ANRs and app hangs grouped into time buckets, so you can see when health changed.

  Try: “Plot crashes and ANRs for the last 14 days.”

  Read · get_app_health_over_time

• Lists error groups

  Returns grouped crashes, non-fatal exceptions and ANRs for an app. You can narrow it by error type and severity.

  Try: “List the top crash groups from the last 7 days.”

  Read · get_errors

• Shows events inside one error group

  Returns individual crash or ANR events for one error group, including the stack trace data.

  Try: “Show me recent events for the top crash group.”

  Read · get_error

• Charts all errors over time

  Returns how many errors occurred per time bucket across all groups, filterable by type and severity.

  Try: “Did errors go up after the last release?”

  Read · get_errors_over_time

• Charts one error group over time

  Returns the occurrence count of a single error group per time bucket.

  Try: “When did this crash first show up and is it growing?”

  Read · get_error_over_time

• Splits one error group by OS, device, version or country

  Shows how a single error group is spread across attributes such as OS, device model, app version and country.

  Try: “Which devices hit this crash the most?”

  Read · get_error_distribution

• Shows the usual path users took before an error

  Returns the most common screen path that led to one error group.

  Try: “What do users do before this crash happens?”

  Read · get_error_common_path

• Lists recent sessions

  Returns sessions for an app, newest first, with filters.

  Try: “Show the latest sessions from users on version 3.2.”

  Read · get_sessions

• Charts session counts over time

  Returns session counts per time bucket.

  Try: “How many sessions did we have each day this month?”

  Read · get_sessions_over_time

• Gets one full session

  Returns a session with all its recorded events, so the assistant can replay what the user did.

  Try: “Walk me through what happened in this session.”

  Read · get_session

• Charts memory use over time

  Returns p50, p90, p95 and p99 memory in KiB by app version and time bucket, for comparing shifts in memory use.

  Try: “Did memory use rise in build 120?”

  Read · get_memory_usage_over_time

• Breaks memory use down by device tier

  Returns p50, p90 and p95 memory per device memory tier with the number of sessions and samples behind each.

  Try: “How does memory use differ on low-end phones?”

  Read · get_memory_usage_by_device_tier

• Lists sessions with high memory use

  Returns a paged list of high-memory sessions. The assistant can then open one with get_session.

  Try: “Find sessions where memory use was highest.”

  Read · get_high_memory_sessions

• Lists bug reports

  Returns bug reports sent from the app, newest first.

  Try: “Show open bug reports from this week.”

  Read · get_bug_reports

• Charts bug report counts

  Returns the number of bug reports per time bucket.

  Try: “Are bug reports increasing?”

  Read · get_bug_reports_over_time

• Gets one bug report

  Returns a single bug report with its full details.

  Try: “Open the most recent bug report.”

  Read · get_bug_report

• Opens or closes a bug report

  Sets a bug report to open or closed. This changes data in your Measure workspace for your whole team.

  Try: “Close bug report 1234, it is fixed in 3.2.1.”

  Write · update_bug_report_status

• Lists top-level trace names

  Returns the names of all root spans recorded for an app, the starting point for performance questions.

  Try: “What traces does the app record?”

  Read · get_root_span_names

• Lists runs of one trace

  Returns individual instances of a root span name with their timing.

  Try: “Show the slowest runs of the checkout trace.”

  Read · get_span_instances

• Charts trace duration over time

  Returns p50, p90, p95 and p99 durations for a span name over time.

  Try: “Has app start-up gotten slower?”

  Read · get_span_metrics_over_time

• Gets a full trace with child spans

  Returns one trace with every child span so the assistant can see where time went.

  Try: “Break down where the time went in this trace.”

  Read · get_trace

• Lists alerts

  Returns the alerts Measure raised for an app, newest first.

  Try: “What alerts fired this week?”

  Read · get_alerts

• Shows user journeys as a screen graph

  Returns the paths users take as a graph of screens and custom events.

  Try: “Which screens do users visit before leaving?”

  Read · get_journey

• Lists journey screens and events with user counts

  Returns each screen or custom event in the journey with its number of users, which you can use to filter get_journey.

  Try: “Which screens have the most users?”

  Read · get_journey_nodes

• Counts users on a journey path

  Returns how many users followed a given path through the journey.

  Try: “How many users went from Home to Checkout?”

  Read · get_journey_path_users

• Lists errors hit on a journey step or path

  Returns the errors users ran into on a chosen screen, event or path.

  Try: “What crashes happen on the payment screen?”

  Read · get_journey_errors

• Lists traces recorded on a journey step or path

  Returns root spans recorded on a journey node or path.

  Try: “Which traces run on the search screen?”

  Read · get_journey_traces

• Lists network calls made on a journey step or path

  Returns the network requests made on a journey node or path.

  Try: “Which requests fire on the login screen?”

  Read · get_journey_network

• Lists sessions that followed a journey step or path

  Returns sessions that went through a journey node or path.

  Try: “Show sessions that reached the cart.”

  Read · get_journey_sessions

• Ranks network endpoints

  Returns the top endpoints by latency, error rate and call frequency.

  Try: “Which endpoints are slowest or fail most?”

  Read · get_network_metrics_trends

• Charts HTTP status classes over time

  Returns 2xx, 3xx, 4xx and 5xx counts over time for an app or one endpoint.

  Try: “Did 5xx errors spike yesterday?”

  Read · get_network_status_codes_over_time

• Charts exact status codes for one endpoint

  Returns counts of each exact HTTP status code for a single endpoint over time.

  Try: “Which status codes does /orders return?”

  Read · get_network_endpoint_status_codes_over_time

• Charts network latency over time

  Returns p50, p90, p95 and p99 latency for an app or one endpoint over time.

  Try: “Has /search latency changed since the last release?”

  Read · get_network_latency_over_time

• Shows when in a session requests happen

  Returns network requests per session grouped by time since the session started, for an app or one endpoint.

  Try: “When in a session do most requests fire?”

  Read · get_network_timeline

• Lists funnels

  Returns an app's funnels, most recently updated first, with steps and conversion windows. You can search by name.

  Try: “List our funnels.”

  Read · get_funnels

• Lists pinned funnels

  Returns the funnels pinned on the funnels page, in their pinned order.

  Try: “Which funnels are pinned?”

  Read · get_pinned_funnels

• Gets one funnel

  Returns a single funnel with its steps, conversion window and pin state.

  Try: “Show the signup funnel definition.”

  Read · get_funnel

• Lists custom events usable as funnel steps

  Returns the custom event names the app has sent, which funnel steps are chosen from.

  Try: “What events can I use in a funnel?”

  Read · get_funnel_event_names

• Shows how many users reach each funnel step

  Returns users per step over a time range and the median time between steps.

  Try: “Where do most users drop off in the signup funnel?”

  Read · get_funnel_conversion

• Charts funnel conversion over time

  Returns conversion by day, week or month, optionally with the builds released in that time.

  Try: “Did conversion change after build 120?”

  Read · get_funnel_trend

• Shows when users enter and convert

  Returns how many users entered a funnel and converted in each hour of the week.

  Try: “What hours do most users convert?”

  Read · get_funnel_cadence

• Lists errors between two funnel steps

  Returns errors users hit while moving from one step to the next, with how many converted or dropped off.

  Try: “What errors block users at the payment step?”

  Read · get_funnel_step_errors

• Lists traces between two funnel steps

  Returns root spans recorded between two steps with their durations for users who converted and who dropped.

  Try: “Are slow traces causing drop-off at step 2?”

  Read · get_funnel_step_traces

• Lists network calls between two funnel steps

  Returns network requests made between two steps with latency for users who converted and who dropped.

  Try: “Which requests are slow for users who drop off?”

  Read · get_funnel_step_network

• Lists sessions between two funnel steps

  Returns sessions of users who converted, or of those who dropped off, between two steps.

  Try: “Show sessions of users who dropped at step 3.”

  Read · get_funnel_step_sessions

• Creates a funnel

  Saves a new funnel from a name, an ordered list of custom events and a conversion window. It is visible to your whole team.

  Try: “Create a funnel for signup_started, email_verified, onboarding_done.”

  Write · create_funnel

• Edits a funnel

  Changes a funnel's name, steps or conversion window for everyone on the team.

  Try: “Add a payment_added step to the signup funnel.”

  Write · update_funnel

• Deletes a funnel

  Permanently removes a funnel. The tool description says a deleted funnel cannot be restored.

  Try: “Delete the old checkout v1 funnel.”

  Delete · delete_funnel

• Pins or unpins a funnel

  Pins a funnel after the others or unpins it. An app can have a limited number of pinned funnels.

  Try: “Pin the signup funnel.”

  Write · set_funnel_pinned

• Reorders pinned funnels

  Sets a new order for all of an app's pinned funnels.

  Try: “Put the checkout funnel first.”

  Write · reorder_pinned_funnels

Watch out for

• delete_funnel removes a funnel for good

  Its description says a deleted funnel cannot be restored. Ask the assistant to confirm before it deletes anything, or sign in with a role that cannot edit funnels. github.com

• Funnel and bug report edits affect the whole team

  create_funnel, update_funnel, set_funnel_pinned, reorder_pinned_funnels and update_bug_report_status change shared data. The docs say funnel edits need a role that can edit funnels in the dashboard. github.com

• ask_question sends your question to Measure's agent

  It runs a Measure-hosted agent that may read raw telemetry with read-only SQL. Do not paste secrets into questions. github.com

Security advisories