Is My Voice Data Safe? A Look at Our Security and Privacy

The short answer: AICHE encrypts your local recordings with AES-256, sends audio to servers via TLS 1.3, processes it through Whisper AI, and immediately deletes the audio file. We keep no logs, no copies, and no activity records.

Dictating sensitive business information, client names, or personal thoughts requires absolute trust that your voice recordings won't be stored, leaked, or used for training AI models.

1. When you record, audio is saved locally in an AES-256 encrypted SQLite database on your device.
2. When processing starts, AICHE sends the encrypted audio through TLS 1.3 (the latest cryptographic protocol) to our servers.

3. Our server receives the audio, sends it to Whisper AI for transcription, receives the text back, and applies any enhancements you enabled.
4. The server immediately deletes the audio file. This happens within 2-3 seconds of receiving your transcription.
5. Only the text transcription returns to your device, where it's stored in your encrypted local history.
6. We log no IP addresses, no timestamps, no user activity. Zero retention means zero data to leak or subpoena.

Heads-up: while audio is deleted immediately, text transcriptions stay in your local encrypted history until you manually delete them. If you dictate sensitive information, clear your history regularly from the History panel.

The pro-tip: the Later queue keeps offline recordings on your device until you process them. These files are AES-256 encrypted, but they exist as audio until processed. For maximum privacy, process your Later queue promptly when internet returns.

The Data Lifecycle

Understanding exactly where your voice data exists at each moment helps you make informed decisions about what to dictate.

Step 1: Recording (Your Device Only)

When you press the hotkey, audio is captured by your microphone and stored in AICHE's local AES-256 encrypted database. At this point, your recording exists only on your device, fully encrypted.

Step 2: Transmission (Encrypted in Transit)

When you stop recording, AICHE sends the encrypted audio to our servers via TLS 1.3 - the most current transport encryption protocol. The connection is authenticated and encrypted end-to-end, meaning no intermediary (ISP, network admin, WiFi operator) can access the audio stream.

Step 3: Processing (Transient Server Presence)

Our server receives the audio, passes it to the Whisper AI transcription engine, receives the text result, and applies any AI enhancements you've configured (Message Ready, Content Organization, Recognition Fix). The audio file exists on the server only during this processing window - typically 2-3 seconds.

Step 4: Deletion (Immediate, Permanent)

The moment transcription completes, the audio file is deleted from the server. Not archived, not moved to cold storage, not retained for any purpose. Deleted. The only data that returns to your device is the text transcription.

Step 5: Local Storage (Your Control)

The text transcription is stored in your local encrypted History. You control this data completely - view it, search it, copy it, or delete it whenever you choose.

What We Don't Collect

Many voice products collect metadata even when they delete audio. AICHE collects none of it:

No IP address logging. We don't record which IP address sent a transcription request. There's no way to trace a transcription back to a network location.

No timestamps. We don't log when you use AICHE. Your usage patterns are your business.

No device fingerprinting. We don't collect hardware IDs, screen resolution, installed software, or any other device characteristics.

No usage analytics on content. We don't analyze what you dictate, what topics you discuss, what languages you speak, or how long your recordings are.

No AI training. Your audio and transcriptions are never used to train or fine-tune any AI model. Your voice data is not a product.

Encryption Details

Local Storage: AES-256

All local data - recordings in the Later queue, text in History, account credentials - is encrypted with AES-256, the same standard used by governments and financial institutions for classified information. The encryption key is derived from your account credentials, so the data is unreadable without your AICHE login.

If someone copies your AICHE database file from your device, they get an encrypted blob that cannot be decrypted without your credentials. There is no master key, no backdoor, and no recovery mechanism that bypasses your encryption.

Transit: TLS 1.3

Audio travels between your device and our servers over TLS 1.3, which provides forward secrecy. Even if a session key were compromised in the future, past transmissions remain encrypted with different keys. This protects against retroactive surveillance.

Compliance and Business Use

For organizations handling sensitive data, AICHE's zero-retention architecture means voice dictation doesn't create new compliance obligations around data storage. Since audio is deleted immediately and text lives only on the user's device, there's no server-side data to audit, subpoena, or breach.

This makes AICHE suitable for environments with strict data handling requirements - legal firms dictating case notes, healthcare professionals recording patient summaries (stored locally and encrypted), and financial services professionals discussing client matters.

For full technical details on our security architecture, visit the AICHE Trust Center.

Result: you dictate 50 voice notes containing client information over two weeks. Each audio file exists for exactly 2-3 seconds on our servers before permanent deletion. No recordings remain anywhere except encrypted on your local device.

Do this now: check the privacy indicator in AICHE's interface to confirm zero retention policy is active, then review our full security documentation at aiche.app/trust for technical details.