Your words stay
on your iPhone.

Privacy policy · Effective October 24, 2026

Orttakey is an iPhone keyboard and companion app made by Olanrewaju Oyinbooke. It turns speech into text using a speech model running on your device. You do not need an account.

Audio and the microphone

You choose when to turn on a voice session. While a session is on, the companion app keeps the microphone active so the keyboard can start dictation from another app. iOS displays its microphone indicator. A rolling buffer holds up to two seconds of recent sound in memory; older idle sound is continuously discarded. Tapping Talk starts a capture and may include a short part of that buffer to avoid losing the first word. Done sends that audio to the on-device speech model. Audio is not saved as an audio file or uploaded to a transcription service. Turn voice off in Orttakey to stop the microphone, or revoke microphone permission in iPhone Settings.

Transcripts and keyboard access

The app saves dictated text and the model’s original transcript in local storage shared with the Orttakey keyboard. This lets you review recent dictations. Orttakey does not build a log of the keys you type, use advertising identifiers, or include advertising or analytics SDKs.

iOS calls the permission for sharing data between a custom keyboard and its companion app “Allow Full Access.” Orttakey uses it to exchange dictation commands, status, and results with the companion app on the same iPhone. It does not use this access to upload your voice or typed text. The ordinary keyboard remains usable without Full Access; voice dictation needs that local bridge.

Downloads and network requests

Setup downloads an approximately 253 MB speech model from Hugging Face. The download host receives the normal technical information associated with a download, such as your IP address and request details, under its privacy policy. Your recordings and transcripts are not part of that download request. After installation, dictation can work offline.

Retention, deletion, and your choices

Saved transcripts remain in the app’s local storage. To remove Orttakey’s local data, delete the app from your iPhone using Delete App (Offload App keeps app data). Device backups are governed by your iOS and iCloud backup settings. You can disable the keyboard, Full Access, and microphone access in iPhone Settings at any time. Removing Full Access disables voice communication with the companion app.

Information you choose to send

If you open a support issue, the details and attachments you choose to provide are handled by GitHub. Please do not include private dictation text unless you want to share it. Apple may separately provide developers with crash reports and App Store/TestFlight information according to your Apple settings and Apple’s policies.

This website

These pages do not use advertising, analytics scripts, tracking pixels, or a sign-up form. The hosting provider processes normal web requests and may retain technical access/security logs. Following a link to Apple, GitHub, or Hugging Face takes you to that service and its policies.

Speech model credits

Fluid Speech Mini by FluidVoice, built on NVIDIA Parakeet TDT 0.6B v3, is provided under CC BY 4.0. Model source and attribution. The Core ML model runs through FluidAudio, licensed under Apache 2.0.

Contact and policy updates

For privacy questions, contact the developer through GitHub, identifying Orttakey in the title. Do not post sensitive information in a public issue. Changes to this policy will appear here with a revised effective date.