Your diary data in the beta

This is a test version. While reliability is being checked, use fictional stories and do not add passwords, documents or information you are not comfortable entrusting to the test organizer.

Where data is stored

Entries are stored in the diary server database, separated by account. The server administrator technically has access to the database and backups. There is no end-to-end encryption.

Voice and AI

Browser voice recognition may use an external browser speech service. Text, rather than audio, is saved by default. AI availability depends on server settings. In the current limited OpenAI pilot, selected text entries and your question are sent to that provider after consent. Audio and photos are not sent for this analysis. The model can be wrong. This is not psychological, medical or investment advice.

Backups and deletion

Backups are created at startup and every six hours of server operation and are accessible to the administrator. You can export entries and delete the current archive under Me → Settings. Deleted entries may remain in backups. Before inviting participants, the organizer must define the retention period and backup deletion procedure.

Drafts on your device

Text drafts and the offline entry queue are stored in the browser without encryption. They are linked to your account but accessible to the device owner. Signing out blocks offline entry until the next sign-in; it does not erase drafts. Clearing site data removes them. Archive API responses are not cached for offline viewing.

Email registration

Website email registration requires an invitation and email verification. A single-use code is sent through webmail.kg and expires in ten minutes. No marketing subscription is enabled. Email is stored with the account; you sign in with username and password.

Google sign-in

Google shares the account identifier, name and verified email. We do not request Gmail or Drive access. To link an existing diary, first sign in to that diary. New website accounts still require an invitation and agreement to the data policy.

Telegram accounts and login

You can create a diary in the bot by reading the policy and choosing Create my diary. Telegram ID, display name and consent are stored. No invitation, Google account or password is required. To connect an existing diary, sign in on the website, open its private bot link and confirm the connection. The bot provides a single-use website login link valid for ten minutes. Never forward it. Confirming creates a seven-day session; use Sign out to end it. Access to your Telegram account can give access to your linked diary. The bot receives only messages sent directly to it.

Telegram photos, voice and disconnection

Text is copied to your archive. New photos are stored as Telegram file identifiers and fetched through an authenticated server route without a permanent archive copy. Availability depends on Telegram; this is not a backup. Previously stored photos are kept. A small copy of your available Telegram avatar may be saved in your profile; your custom photo is not overwritten. Telegram voice is transcribed using local Whisper on the diary server. Audio is processed in memory, not kept in the archive or backups. Original messages and audio remain in Telegram until you delete them there. Choose Disconnect Telegram or send /unlink to stop capture. Deleting the archive also disconnects Telegram. Disconnecting alone does not delete saved entries. Refresh the website to see new entries.

Recovery and participation

Website registration needs an invitation; bot registration does not. Keep your personal recovery code separately if one was issued: it can reset your password. Sign out using the account bar. Before external testing, the organizer must supply their name, a support contact and a data retention period. This page describes technical behavior and does not replace the necessary participation terms.

Back to the diary · Diary AI