Privacy

e1-4 is built and operated by Earth Life-Forms Ltd, the data controller. Your voice is yours. This page explains what we store, how your voice signs you in, and how to take everything back. Voice policy version voice-v1; training objection policy version train-optout-v1. Last updated 2026-10-18.

Sign-in

You sign in with your voice — there is no typing and no e-mail sign-in. A passkey on your device is an optional backup way in.

What we store

Voice Stream audio entries, their timestamps, and their transcripts when a transcriber is configured.
Voice notes you send in Talk conversations.
Infinity Chalkboard boards you save while signed in.
Share links you create, and whether they have been revoked.
Your contacts, and your profile handle and avatar if you set them.
Your earth life-hive bio and posts. Posts are public to anyone who opens your profile link, and are deleted when you delete them or your account.
A sealed voiceprint and consent records (below).

Voice authentication & voice data (voice-v1)

When you speak to e1-4, the app computes a voiceprint: numeric features derived from your voice — not the audio itself. The voiceprint is sealed with AES-256-GCM under a server-held key, bound to your account, and used only to recognise you at sign-in and to check that stream entries were spoken by the stream's owner. What you say to sign in is also saved as an entry in your Voice Stream, once your consent is on file. Every recording is encrypted with AES-256-GCM before it is stored, and only e1-4's servers hold the key. When liveness checking is enabled, the sign-in also asks you to speak digits shown on screen to prove a live person is speaking.

Consent

A voiceprint is biometric data under GDPR Art. 9, the Illinois Biometric Information Privacy Act (BIPA), Texas CUBI, and Washington law. We collect it only with your explicit consent. When you agree, we record the policy version (voice-v1), the time, a hashed network address, your device string, and a short voice recording you made when you agreed, kept as proof that you agreed. Consent is checked before any new voice saving; withdrawing it stops all new saving immediately.

Sharing & processors

Your data is not sold, not shared for advertising, and not used to train third-party models. Processors: Microsoft Azure (database and file storage). Speech-to-text runs on your own device, or through e1-4's own self-hosted Da Vinci service when that is enabled.

Training Da Vinci

By default, the text of your conversations with Da Vinci — what you say, its reply, and the text of what Da Vinci hears you say — is copied to a training store run by Earth Life-Forms Ltd and used to train e1-4's own Da Vinci models. Our legal basis is legitimate interest (UK and EU GDPR Art. 6(1)(f)): improving the assistant you are talking to. Voice recordings, voiceprints and your saved stream are never used for training, and nothing is used to train anyone else's models. You can object at any time with the training glyph on your profile: your past conversations are deleted from the training store and no new ones are added. Deleting your account does the same. A model trained before you objected cannot unlearn what it saw, but it is never trained on your data again.

Retention & deletion

Everything is kept until you delete it. Deleting a single entry or your whole stream removes the stored audio files. Deleting your account destroys the voiceprint, all audio and every record — including your consent records and the proof recording. You can delete immediately, or schedule deletion: it takes effect after a 30-day cool-off (during which the deletion can be cancelled from your profile) and is then carried out by an automated job.

Withdrawing consent

Tap the withdraw glyph on your profile to withdraw voice consent at any time; all new voice saving stops right away. Delete your account to erase the voiceprint itself.

Your rights (GDPR / CCPA)

Access & portability: download everything from your profile as JSON.
Erasure: deleting a segment, your whole stream, or your account hard-deletes database rows and stored audio/avatar files.
Withdrawal: voice consent can be withdrawn at any time from your profile.
Objection: stop Da Vinci training on your conversations from your profile (see Training Da Vinci).
No sale of personal information: we do not sell or share personal information for cross-context behavioural advertising.

e1-4 is not for children under 13 (under 16 in the EU).