Agent API

Read-only analytics access for coding agents: the same numbers the dashboard renders, over REST and MCP.

Other guides: quickstart.md · events.md · revenuecat.md · apps.md · apple-ads.md

Auth

Every request:

Authorization: Bearer <key>

(x-api-key: <key> also works.) Mint keys at dashboard → gear icon → agent api tab. A key (ap_sk_…) belongs to your account, is shown once (stored hashed), and is scoped at creation — all your apps (new apps included automatically) or a selected list; edit the scope or revoke it there anytime. Every tool answers only within the scope: list_apps returns the scoped set, app: "all" aggregates exactly that set. Keys are read-only, distinct from the per-app ap_pk_ ingest key. Never commit or print one — use an environment variable.

REST

GET with query params, or POST with a JSON body of the same args, against https://activationpal.com/api/v1/<tool>:

curl -s "https://activationpal.com/api/v1/get_stats?app=myapp&range=7d" \
  -H "Authorization: Bearer $AP_KEY"

curl -s https://activationpal.com/api/v1/get_events \
  -H "Authorization: Bearer $AP_KEY" -H "Content-Type: application/json" \
  -d '{"app": "myapp", "name": "paywall_shown", "limit": 100}'

Responses: { "ok": true, ... } or { "ok": false, "error": "..." }. GET /api/v1 (authed) lists every tool with its JSON schema.

MCP

Streamable-HTTP, stateless, same tools and key: https://activationpal.com/api/mcp

claude mcp add --transport http activationpal https://activationpal.com/api/mcp \
  --header "Authorization: Bearer <key>"

Tools

`list_apps`

No params. Returns every tracked app: id (the app value in every other tool), display name, IANA timezone, App Store adam_id.

`get_stats`

The full analytics snapshot — the exact object the web dashboard renders.

Param	Type	Notes
`app`	string, required	App id from `list_apps`, or `all`
`range`	string	`today`, `yesterday`, `week` (calendar week, Mon-first), `7d` (default), `30d`, `month`
`tz`	string	IANA override; default: the app's own timezone (apps.md) — every bucket and daily cut runs in it
`users`	boolean	Include the per-device rollup (long; default false)

Returns: headline installs / onboarded / devices / app-opens / purchases / ARPU / conversion + paywall-conversion rates (install-cohort) / average + median session length with previous-period values; gross revenue + mix (new subs / renewals / one-time); pricing-tier splits with per-country detail; time-bucketed chart series; cohort onboarding funnel (steps, answer splits, completion, paywall shown → plan → purchased with plan splits); paywall placements; acquisition (Apple Ads vs organic via AdServices); Apple Search Ads spend / taps / CPI; RevenueCat health (trials, conversions, churn reasons, by-product); geo by country and city; device / OS / app-version / language breakdowns; custom event counts; per-app distribution in all mode.

`get_events`

Recent raw events, newest first. Each row: name, ts, device_id, user_id, country, city, source, env, props.

Param	Type	Notes
`app`	string, required	App id, or `all`
`limit`	number	1–500, default 50
`name`	string	Exact event name — `paywall_shown`, `rc_renewal`, any custom event
`device_id`	string	Only this device's events
`source`	string	`sdk` or `revenuecat`
`env`	string	`release`, `testflight`, `debug`
`since`	string	ISO timestamp — only events at/after it

`get_user`

One device's full timeline for an app, oldest first: every SDK event it sent plus the RevenueCat events of every user id it reported — replays a single user's journey. Params, both required: app, device_id (from get_stats with users: true, or get_events).