Privacy

The short version

Deadlock Meta is an unofficial fan site. Public stats can be browsed without an account. Optional Replay Lab coaching uses a separate, private app with accounts and uploaded replays. No account is required to search public player stats, and a public profile lookup does not verify that you own that profile. We do not run our own analytics or tracking.

Public voting cookie

If you vote in the tier list, we set a single first-party cookie named “dm_voter”. It holds a random, anonymous id — not your name, not your Steam account, nothing that identifies you. Its only job is to keep one browser from voting twice on the same hero and to show your own picks when you come back. If you never vote, we never set it. Clearing your cookies removes it (and disconnects you from your past votes).

Player searches

When you (or anyone) look up a player, we fetch that player's already-public Deadlock data — their Steam display name and avatar, and their match results (hero, win/loss, K/D/A, and similar per-match stats) — from the community API at deadlock-api.com, and cache the most recent slice of it so repeat visits are fast and we don't hammer that API. This is public game data, the same information any Deadlock stats site shows; it is not private and we don't attach it to you, the visitor. If you're a player and want your cached data removed, email us (below) and we'll delete it.

Where the data comes from

Match stats and hero/item images come from deadlock-api.com, a free community project (which in turn draws on Valve's public game APIs and player-submitted match data). We don't control what that API collects; its own practices are its own. Deadlock and all related names, images, and assets are property of Valve Corporation — this is an unofficial fan site, not affiliated with or endorsed by Valve.

Private Replay Lab accounts and uploads

When the coaching app is configured and you create an account, we store your email, optional display name, a password hash and revocable sign-in sessions. The app sets a separate HttpOnly, host-only sign-in cookie; the anonymous voting cookie is never used to authenticate you. Email verification and self-service password recovery are not yet available in this pilot.

An uploaded replay can contain other participants' names, identities and gameplay. We extract one shared match ledger and player-specific reports. Reports are accessible only through authenticated ownership checks, not through the public player search. Equivalent uploads can reuse a processing cache, but another account must submit the replay itself before it can read that match's reports. Uploading does not publish a report or attach it to a public profile.

When AI-written critique is enabled, the facts our parser extracts from your replay — player identifiers, heroes, item purchases, deaths, objective timings and match statistics — are sent to Anthropic (the Claude API) to write the critique. The replay file itself is never sent. Anthropic retains API inputs under its own policy (currently up to 30 days) and does not train on them. Every report says whether it includes AI-written critique or only recorded evidence.

Managed raw replay copies expire after seven days by default once processing is finished. Extracted ledgers and reports currently remain stored without an automatic expiry; sign-in and billing records are also retained for account operation. Contact us to request account or report deletion. Deleting one account must not remove another account's access to a legitimately shared processing result.

Subscriptions

When billing is configured, Stripe hosts checkout and subscription management. We retain customer/subscription identifiers, payment-confirmed access state and billing-period information. We do not store card numbers. A checkout return page does not activate paid access; signed payment and subscription notifications do. Taxes and final charges are shown by checkout when configured.

Advertising

Public stats pages have optional display-ad integrations, off by default. Before activating one, we must name the partner here and configure and test the required consent handling. Advertising partners may process browsing data and set cookies when permitted. Private coaching runs in a separate ad-free document: no ad scripts or advertising trackers are loaded there, including sign-in and error pages.

What we don't do

We don't sell your replay data, publish private reports by default, or use public profile searches as proof of account ownership. We don't fingerprint you or track you across other sites. Optional coaching accounts and subscriptions are separate from anonymous public browsing.

Your choices

You can clear or block the dm_voter cookie in your browser at any time (voting just won't remember you). When ads are on, you can decline personalised ads through the consent prompt. To have a cached player profile removed, contact us.

Contact & changes

Questions, or want data removed? Email [email protected]. We'll update this page if what we store changes; the date below reflects the last revision.