Privacy policy

What we process

• Text you choose to rewrite, explain, summarize or have read aloud. When you trigger an action, the selected or pasted text, the action and style you chose and any instruction you typed are sent over an encrypted connection to our server. For rewrite, explain and summarize the server forwards the text to Anthropic's Claude API and streams the result back to you. For read aloud the server forwards the text to ElevenLabs, which turns it into speech that is streamed back to you. Our servers do not store the text, the result or the audio, and they are not written to logs.
• Audio you dictate. When you dictate, the recording from your microphone is sent to our server, which forwards it to ElevenLabs for transcription. The transcript is then sent to Anthropic's Claude API to be turned into clean written text. The recording is not stored by us or by ElevenLabs beyond the time needed to transcribe it. Your microphone is only active while you hold or toggle the dictate button.
• Anonymous device record. On first launch, or on first use of the web reader, a random device token is created (in the app's Keychain or credential store, or in your browser's local storage). We store a hashed version of it together with the platform, the app version, the plan, a salted hash of your IP address and timestamps. It is not linked to your name or email.
• Usage records. For each action we store the time, the action and style, the number of characters, tokens or seconds of audio, the duration and whether it succeeded. We use these to enforce the free daily quotas, to understand load and to pay for the AI services. Usage records are deleted after 12 months.
• Reader settings. The web reader keeps your text size, language and speed choices in your browser's local storage. They never leave your device.
• Waitlist email. If you join the Pro waitlist we store your email address until Pro launches or you ask us to remove it.

What we don't do

• We don't sell or share your data with advertisers.
• We don't use your text or voice to train AI models, and neither do Anthropic or ElevenLabs under their API terms.
• We don't read your screen, clipboard or microphone beyond the text you explicitly select and the moments you dictate.

Sub-processors

Anthropic, PBC processes text for rewrite, explain, summarize and the polishing step of dictation under their commercial API terms. Those terms include a limited retention period for abuse monitoring (currently up to 30 days) and no training on API data.

ElevenLabs, Inc. processes text for read aloud (text to speech) and audio for dictation (speech to text) under their API terms with zero data retention. Recordings and generated audio are not stored after the request completes and are not used for training.

Our hosting provider stores the database described above in the EU.

Own API key mode

If you enter your own Anthropic API key in the app's Settings, the text features talk to Anthropic directly from your computer. Your text never reaches our servers at all, and no usage records are created for those actions. Read aloud and dictation still go through our server.

Legal basis and your rights

We process data to provide the service you request (performance of a contract) and to protect it from abuse (legitimate interest). Under the GDPR you may request access, correction or deletion of data we hold about you. Because device records are anonymous, include the device ID shown in the app's Settings.

Email privacy@zimplfy.io. Addresses at rewrite4me.io keep working.

Changes

We will post changes here and update the date above. Significant changes are announced in the app.