✦ Now with returning players and D1, D7, D30 retention
☰
All featuresAttributionRetentionAnalytics OverviewIntegrationsPricingDocumentationComparisonsFree toolsBlogGlossary
Privacy and definitions

Identity and collection choices

Anonymous ID

The browser SDK creates a random UUID as the anonymous_id. It stores the ID in a first-party cookie. If cookies are not available, it uses localStorage, then memory. The ID persists across days in the same browser, so returning visitors and cross-day journeys are measured. It is not derived from IP address or user agent.

The SDK also keeps a client session_id. A new session ID starts after 30 minutes of inactivity.

Links from a game

When a game opens your website with the Godot SDK's link_url() or the Unity SDK's LinkUrl(), the link carries the game's random anonymous ID for 10 minutes. The website SDK removes it from the address bar and sends it once, so reports count the game install and the browser as one person. It is not personal data. In consent mode, nothing is sent before consent.

Optional consent mode

Set requireConsent: true (or data-require-consent="true" on the snippet) to start in consent mode. Before consent, the SDK stores nothing in the browser and sends events with no anonymous ID, session ID, or user ID. Those events count toward event totals but not visitors. Call setConsent(true) to start the persistent identity. Call setConsent(false) to clear identity storage and strip identity from queued events. optOut() stops tracking for that browser.

Visitor counts and retention

Unique visitors are distinct anonymous IDs. A person returning tomorrow in the same browser counts once and appears as a returning visitor. Day and week cohort retention (D1, D7, D30) use these IDs and explicit user IDs. A cleared cookie, private window, or second device creates another anonymous ID.

Accounts and traits

Call identify(userId, traits) after login. The SDK sends an identify event and adds user_id to later events. Scalar traits merge by key in Tinybird. Call reset() on logout to forget the user ID, anonymous ID, and session.

Server-side events

Backends can send track and identify events to POST /v0/events with channel: 'server' and the project's Secret key in Authorization: Bearer. Keep the Secret key on your server only. Each event needs an anonymous_id or a user_id.

Revenue identity

Revenue comes from the store where you sell. Steam data comes from the Steam UTM Analytics report and, with a Steamworks Financial API key (the automatic Steam sync is in early access), a daily sync of wishlists and sales. Steam reports totals, so Steam sales are counted by source and campaign.

An itch.io API key syncs purchases and donations as purchase events with the channel itch. They are matched to a player by an email trait that your backend sends with identify. An owner or admin can turn email matching off.

If your game also sells on its own website, signed Stripe or Paddle events supply invoices, refunds, and subscription snapshots. Link them with provider metadata user_id or identify traits stripe_customer_id / paddle_customer_id. Email similarity is not used to guess a Stripe or Paddle match. Unmatched revenue remains visible; currencies are kept separate.

Event data

The collector does not store raw IP addresses or raw user agents. It derives only country, device, OS, and browser. Country comes from an in-memory lookup of the visitor's IP address: Cloudflare's published geofeed for Cloudflare addresses such as WARP, then the DB-IP country database (or an explicit country header from your proxy); the address is then discarded and never stored or logged. IP Geolocation by DB-IP. Page events include origin, path, title, external referrer origin, sanitized UTMs and supported ad click IDs. Other query strings and fragments are removed. Requests from known bots and crawlers are kept with the device label bot and excluded from reports. Custom properties and traits are supplied by site owners.

Consent and operational data

The SDK uses a first-party cookie for the anonymous ID. Site owners decide whether their sites need consent and can use the consent mode when they do. Review infrastructure logs, queue retention, custom properties, and deletion policies separately.

Installation and consent examples · How players are counted

Frequently asked questions

Does Manalens use cookies?

Yes, one first-party cookie. The browser SDK stores a random anonymous ID in it, with localStorage and then memory as fallbacks. The ID is not derived from the IP address or the user agent. In consent mode, the SDK stores nothing in the browser until the visitor consents.

Are IP addresses stored?

No. The collector does not store raw IP addresses or raw user agents. It looks up the country from the IP address in memory, with the Cloudflare geofeed and then the DB-IP country database, and then it discards the address. It keeps only country, device, OS and browser.

How does consent mode work?

Set requireConsent to true to start in consent mode. Before consent, events carry no anonymous ID, session ID or user ID, so they count toward event totals but not visitors. Call setConsent(true) to start the persistent identity, or setConsent(false) to clear identity storage.

How is one player counted across the game and the website?

When a game opens your website with the Godot link_url() or the Unity LinkUrl() call, the link carries the game's random anonymous ID for 10 minutes. The website SDK sends it once and removes it from the address bar, so reports count the install and the browser as one person.