ChatGPT sign-in supplies a site-specific user ID, email and optional name. EpicHear stores saved show IDs and listening positions against your user ID in its hosted database. Wallet records, when present, are also associated with that ID. In My library, you can download your account records and clear saved stories and listening history. This does not delete your ChatGPT account.
Browser storage
Saved creator stories are stored privately in your account. An unfinished working copy and older drafts may be kept in browser storage. Playback positions are also stored on this device to help you resume. Clearing library history removes playback positions from this browser and your account; pause other devices first.
Payments
No payment details are collected by EpicHear. Coin purchases are unavailable.
Hosting
ChatGPT Sites provides hosting and sign-in. Its own privacy terms apply to those services.
Creator waitlist
If you subscribe, we store your email, signup date and consent version to send creator programme updates and the submission-opening announcement. This is separate from analytics consent. No story idea is collected at signup. Your email is visible only to the site owner, not displayed publicly, and kept until you ask to leave or the programme ends. To unsubscribe or request deletion, email info@epichear.com.
Campaign attribution
With optional analytics enabled, we store the first traffic source, campaign medium and campaign label for this browser tab in session storage and attach them to analytics events. Campaign labels are limited to 64 letters, digits, underscores or hyphens; please never put personal information in campaign links. Events are retained for up to 90 days. Rejecting analytics clears the session attribution. This does not change access to audio.
Analytics, session replay and personalized story picks can be controlled separately in privacy preferences. Outside the US, optional analytics and replay stay off until you choose them. US visitors see no automatic consent popup: analytics and masked public-page replay are enabled unless opted out. Existing choices, browser exclusions, Global Privacy Control and Do Not Track take priority. Unknown locations use the consent-first experience. All visitors outside the US receive the privacy prompt. Listening works with all tracking off. Choices are saved in this browser’s local storage.
Aggregate public traffic
We count public HTML page requests in separate daily totals by page category, broad country, device category and recognized referral source. These counters store no IP address, raw user agent, raw referrer, visitor ID or per-request record and cannot reconstruct a visitor journey. They are kept for up to 90 days and are not shared with Clarity. Known bots, owner requests, Global Privacy Control, Do Not Track and browser exclusions are excluded. Choose Keep off in privacy preferences to opt out of future counts; a 90-day first-party preference cookie records that choice. These totals are requests, not unique people. Choosing analytics only does not enable replay.
Optional analytics
When analytics is enabled, EpicHear records page-view counts, show clicks, audio starts, listening time, playback positions in seconds at periodic checkpoints and stops, unique chapter coverage, buffering and playback errors, sign-in and reward-claim clicks, story cards at least half visible for one second, their placement and cover version, recommendation clicks and preference feedback, along with timestamps, a playback-attempt ID, an ordered event timestamp and a session ID that persists across reloads until 30 minutes of inactivity and a random browser ID. The browser ID expires after 90 days. With analytics enabled and after sign-in, it can be linked to a server-generated pseudonymous account key to measure returning listeners and cross-device use. We do not automatically assign a shared browser’s anonymous history to one account when several accounts have used it. Events may include approximate country supplied by our hosting network, device category, browser family and browser language. Country can be inaccurate or unavailable; a phone’s exact model and your income are not collected or inferred.
Custom events also include a fixed public-page category (such as Home, Story page or a named journal guide) and a broad source label (such as Google, ChatGPT or Product Hunt). Source labels are derived in the browser from a recognized referral domain or an allowlisted utm_source value. Raw referral URLs, domains, query strings and arbitrary campaign values are not sent to our event service. Direct / unknown includes visits whose source cannot be determined. Saved choices are respected. Regional defaults apply only where no choice has been saved. Our custom event records do not contain your name, email, IP address, raw User-Agent or exact page addresses. The report is owner-only. Custom events older than 90 days are removed during subsequent analytics submissions; deletion can be delayed while no submissions arrive. Older records may lack the new fields. General page tracking excludes account, studio, sign-in and admin routes. Named sign-in clicks, wallet opens, reward-claim attempts and ongoing audio-player events may be recorded with analytics enabled on account and sign-in pages. Studio and admin pages remain excluded. An unsent event queue is stored on this device for up to 24 hours, bounded to 250 events per tab, for retries after connection failures. It is cleared when analytics is turned off. Turning analytics off removes the browser ID and stops future optional analytics; previously submitted events remain subject to the cleanup process described above.
Confirmed service actions
Welcome coins, daily rewards, timed waits and episode unlocks are recorded as service transactions whether or not optional analytics is enabled. For up to 90 days, an owner-only operational report summarizes those confirmed actions using a pseudonymous account key; owner actions are excluded from its audience totals. These are separate from consented browsing events. Transactions remain in your wallet history according to the service’s account-record retention. Named click events are never treated as proof a reward or unlock succeeded.
Service security
Short-lived request counters tied to your site account limit repeated writes. Expired counter rows are cleaned up during subsequent write activity. These counters do not use IP addresses or device fingerprints.
Cloudflare Web Analytics
When analytics is enabled, we also load the Cloudflare Web Analytics beacon to measure traffic and page performance. This makes requests to Cloudflare and is separate from EpicHear’s custom event records. Cloudflare’s service receives technical request information, including page URLs and browser/network information. Avoid putting personal information in URLs. See Cloudflare’s privacy policy. When analytics is turned off, the page may reload to unload the beacon.
Microsoft Clarity heatmaps and session replay
Session replay has its own control. A saved analytics-only choice does not enable it. When replay is enabled through your choice or the US regional default, Microsoft Clarity receives public-page URLs and technical device information and reconstructs clicks, scrolling and page interactions to produce heatmaps and session replays. Form fields, search content and episode discussions are masked. Account, sign-in, studio and owner pages are excluded, as are recognized owner accounts and excluded test browsers. We do not send your account name, email or account ID to Clarity. Advertising storage is denied. You can turn replay off in privacy preferences; the page reloads to unload the recorder. Standard recordings are available in Clarity for 30 days; heatmaps and some retained recordings may remain for up to nine months. See Microsoft’s privacy statement.
Personalized story picks
If you enable personalized picks, story choices, saves, listening time, episode completions and “More like this” or “Not for me” feedback build a taste profile in this browser. Interests expire after 90 days without activity on that story. This profile stays on this device and does not sync across devices or accounts. Turning personalization off deletes it. You can also use “Reset taste” on the home page. Recommendations match story genres and themes; longer listening and feedback carry more weight than opening a story. We may also use your approximate network country and device category to select group recommendations from aggregated listening. Each story must have at least ten distinct consenting browser IDs in a group before it is eligible; otherwise we use broader groups or editorial picks. Your own story interests take priority. Country and phone type are not treated as evidence of income or personal taste. The group lookup does not create an individual profile on our server.
Aggregate audio delivery counts
When an episode file is served, we count qualifying requests by UTC day, book and episode without waiting for optional analytics consent. These counters contain no visitor or session identifier, IP address, account, device, country, referrer or individual timestamp. We do not join them to listening profiles. Requests starting at byte zero and delivering at least 1 KB count; tiny probes, later seek ranges, HEAD requests, failed or locked requests and recognized bots do not. Caches, preloading and retries affect totals, so these are not unique listeners or confirmed plays. Counts older than 90 days are cleaned up on subsequent qualifying requests; cleanup may be delayed during inactivity. Hosting providers still process technical request data to deliver the service. We also honor Do Not Track and Global Privacy Control for these counts.
Exclude this browser from measurement
Use this for your own testing, or to opt out of aggregate audio delivery counts. It also stops optional analytics in this browser. Apply it separately on each device. This does not remove earlier records or hosting security logs.
Episode discussions
Comments and the nickname you choose are public. Your account identifier is used privately to let you delete your comments; it is not displayed. Comments remain until deleted. You can delete your own comment from its discussion and export your comments from My library. Report abuse using the Report link.
