Engageable Analytics — Claude Code MCP server: GA4, Mixpanel, PostHog
One MCP server in place of three integrations: your agent pulls numbers from GA4, Mixpanel and PostHog and runs the analysis itself.
claude mcp add --env ANTHROPIC_API_KEY=<ANTHROPIC_API_KEY> --env POSTHOG_API_KEY=<POSTHOG_API_KEY> --env POSTHOG_PROJECT_ID=<POSTHOG_PROJECT_ID> --env GA4_CREDENTIALS_JSON=<GA4_CREDENTIALS_JSON> --env GA4_PROPERTY_ID=<GA4_PROPERTY_ID> --transport stdio engageable-analytics -- uvx engageableWhat Engageable Analytics does
Engageable Analytics plugs into Claude, Codex, Cursor and other MCP clients, exposing 9 tools. analyze_trends handles time series with change points and anomaly detection, compare_segments runs A/B comparisons with significance, analyze_funnel walks a multi-step funnel and reports drop-off, and analyze_retention builds D1/D7/D30 cohort curves. The ask tool takes a plain question and picks which of the others to call.
Every tool is its own pipeline: parse the request, hit the right connector (ga4_query, posthog_query, mixpanel_query), compute, return. Tabular results come back as CSV rather than JSON, capped at a 1000-cell budget so the context window stays workable. Transport is stdio, or SSE at http://localhost:8080/sse under Docker.
You bring your own Anthropic key — the natural-language routing layer calls an LLM, so the server will not start without ANTHROPIC_API_KEY. Source credentials go in as environment variables or get saved by configure_source to ~/.engageable/credentials.json. PostHog wants an API key plus a project ID, Mixpanel a service account and project ID, GA4 a service account JSON and a property ID.
What it helps with
- build a conversion funnel with drop-off per step
- spot spikes and drops in traffic
- compare two user cohorts side by side
- measure D1/D7/D30 retention
- ask about metrics in plain language
Example prompts for the agent
- Show signups for the last 30 days and flag any change points
- Compare purchase conversion before and after the new checkout shipped
- Pull retention curves for March cohorts
How to install Engageable Analytics
Platform CLI
claude mcp add --env ANTHROPIC_API_KEY=<ANTHROPIC_API_KEY> --env POSTHOG_API_KEY=<POSTHOG_API_KEY> --env POSTHOG_PROJECT_ID=<POSTHOG_PROJECT_ID> --env GA4_CREDENTIALS_JSON=<GA4_CREDENTIALS_JSON> --env GA4_PROPERTY_ID=<GA4_PROPERTY_ID> --transport stdio engageable-analytics -- uvx engageableReplace <…> with your own values. 3 more environment variables are in the server README. uvx runs the command named after the package; if the README names another one, use uvx --from engageable <command>
Platform CLI
codex mcp add engageable-analytics --env ANTHROPIC_API_KEY=<ANTHROPIC_API_KEY> --env POSTHOG_API_KEY=<POSTHOG_API_KEY> --env POSTHOG_PROJECT_ID=<POSTHOG_PROJECT_ID> --env GA4_CREDENTIALS_JSON=<GA4_CREDENTIALS_JSON> --env GA4_PROPERTY_ID=<GA4_PROPERTY_ID> -- uvx engageableReplace <…> with your own values. 3 more environment variables are in the server README. uvx runs the command named after the package; if the README names another one, use uvx --from engageable <command>
Config file
{
"mcpServers": {
"engageable-analytics": {
"command": "uvx",
"args": [
"engageable"
],
"env": {
"ANTHROPIC_API_KEY": "<ANTHROPIC_API_KEY>",
"POSTHOG_API_KEY": "<POSTHOG_API_KEY>",
"POSTHOG_PROJECT_ID": "<POSTHOG_PROJECT_ID>",
"GA4_CREDENTIALS_JSON": "<GA4_CREDENTIALS_JSON>",
"GA4_PROPERTY_ID": "<GA4_PROPERTY_ID>"
}
}
}
}Replace <…> with your own values. 3 more environment variables are in the server README. uvx runs the command named after the package; if the README names another one, use uvx --from engageable <command>
Config file
{
"servers": {
"engageable-analytics": {
"type": "stdio",
"command": "uvx",
"args": [
"engageable"
],
"env": {
"ANTHROPIC_API_KEY": "<ANTHROPIC_API_KEY>",
"POSTHOG_API_KEY": "<POSTHOG_API_KEY>",
"POSTHOG_PROJECT_ID": "<POSTHOG_PROJECT_ID>",
"GA4_CREDENTIALS_JSON": "<GA4_CREDENTIALS_JSON>",
"GA4_PROPERTY_ID": "<GA4_PROPERTY_ID>"
}
}
}
}Replace <…> with your own values. 3 more environment variables are in the server README. uvx runs the command named after the package; if the README names another one, use uvx --from engageable <command>
Platform CLI
gemini mcp add -e ANTHROPIC_API_KEY=<ANTHROPIC_API_KEY> -e POSTHOG_API_KEY=<POSTHOG_API_KEY> -e POSTHOG_PROJECT_ID=<POSTHOG_PROJECT_ID> -e GA4_CREDENTIALS_JSON=<GA4_CREDENTIALS_JSON> -e GA4_PROPERTY_ID=<GA4_PROPERTY_ID> engageable-analytics uvx engageableReplace <…> with your own values. 3 more environment variables are in the server README. uvx runs the command named after the package; if the README names another one, use uvx --from engageable <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Engageable Analytics
How do I install Engageable Analytics in Claude Code?
Run `claude mcp add analytics -e ANTHROPIC_API_KEY=<ANTHROPIC_API_KEY> -e POSTHOG_API_KEY=<POSTHOG_API_KEY> -e GA4_CREDENTIALS_JSON=<GA4_CREDENTIALS_JSON> -e MIXPANEL_SERVICE_ACCOUNT_SECRET=<MIXPANEL_SERVICE_ACCOUNT_SECRET> -- uvx engageable`.
What does Engageable Analytics do?
It hands your agent 9 analytics tools covering trends, segment comparison, anomaly detection, funnels, retention and cohorts. Data comes from GA4, Mixpanel and PostHog through a single MCP server.
What do I need before running it?
An Anthropic key, since the natural-language layer calls an LLM, plus credentials for your sources: a PostHog API key and project ID, a Mixpanel service account, or a GA4 service account JSON and property ID.
