This website
The FearKeeper site is a folder of static files. It does not run scripts, set cookies, or embed a form, font, or analytics service. Joining the waitlist opens your own email app and addresses a message to hello@fearkeeper.app. That message is ordinary email you send. This page cannot see whether you sent it.
The pages name fearkeeper.app as the canonical address. fearkeeper.com is intended for the same site.
The app does not have an account
FearKeeper does not ask you to sign in. Nothing in this repository sends campaign notes, recordings, or transcripts to a server run for FearKeeper, and there is no analytics library in the project.
What stays on the device
- Campaigns, sessions, the party, countdowns, the codex, and transcripts are stored in a local database on the device. iCloud sync is turned off. The app’s entitlements do not include an iCloud container.
- The database models are written so they could be synced later. A comment in the project says to switch that on only after an iCloud container is configured. That switch is not on. If it is turned on later, this page has to say so before it can still be trusted.
- Session audio is a file in the app’s Application Support folder, in a Recordings directory, not a row inside the database. The app writes the file while you record.
- A transcript is text stored with that recording. Paragraphs can carry a timestamp.
- Deleting a recording deletes the audio file and the transcript kept with it.
Recording and transcription
Recording uses the microphone. The system asks permission first. On iPhone and iPad, a recording can continue while the screen is locked. You can pause and resume. The app records one session at a time.
Transcription uses Apple’s Speech framework on the device. The in-app description tells you the recording is turned into text on that device. Where live transcription is available, the transcript grows while you record. Otherwise the saved audio is transcribed after you stop.
The first time a speech model is needed, the system may download and install that model onto the device. That is a download of the model, not an upload of your audio by FearKeeper. FearKeeper’s own code does not contain a request that sends the recording or the transcript anywhere.
Transcription follows the device’s language when Apple supports it, and otherwise tries US English. It can be unavailable on a given device. The app says so when it cannot transcribe.
You can copy a transcript, share that transcript, or share the audio file, through the system share sheet. Any of that leaves the device only because you sent it. The app has no playback screen for the recording; the file is what transcription and sharing use.
If the app quits during a recording, it tries on the next launch to keep the audio already written, or the transcript saved so far, and to drop audio files that no recording still points at.
What a finished policy still needs
Before this draft is replaced, Hero Huddle LLC still has to decide how waitlist email is stored and for how long, who can read it, and what the notice should say if iCloud sync, an account, or any upload is added. This page does not set those rules, and it does not describe encryption beyond what the device’s normal app storage already does.