Your voice. Your data. On your device or self-hosted.
No limits · Private · Encrypted · 99 languages
Fast and accurate speech to text iOS keyboard that works in any app. No word limits, no daily caps.
No weekly caps, no word restrictions, no catch. Dictate as much as you want, whenever you want.
Speech recognition that runs locally on your iPhone. No internet required. Audio never leaves your device.
AES-256 end-to-end text encryption with X25519 key exchange. Zero analytics, zero tracking, zero data collection.
Automatic multilingual speech recognition. Speak in your language, or translate to another - Diction handles both.
Turns natural speech into polished text. Automatic context-aware grammar and formatting upgrade.
Cloud transcription with zero setup. Highest accuracy, lowest latency. Audio processed and immediately discarded.
On-device and self-hosted modes are completely free with basic models. No word limits, no daily caps. Pro unlocks premium on-device models for better accuracy, plus Diction One cloud. Free trial included.
Diction uses Whisper-based models that are significantly more accurate. No session time limits, no word caps. Works identically across all apps. Choose between on-device, cloud, or your own server.
Yes. On-device mode works without internet once the model is downloaded. Cloud and self-hosted modes require network access.
99 languages via Whisper. On-device base model handles most languages well. Cloud and self-hosted modes use larger models for even better accuracy across all supported languages.
Never. On-device mode processes audio in memory and discards it immediately. Self-hosted mode sends audio only to your server - we have no access. Diction One cloud processes and discards. No recordings retained, no model training.
After transcription, Diction can optionally clean up your text - removing filler words, fixing grammar, and polishing the result. Only the text is sent to the AI, never the audio. Off by default.
Open the app, grant microphone permission, add Diction as a keyboard in iOS Settings, enable Full Access, and start dictating. Under a minute from download to first transcription. Detailed steps here.
You run a Whisper speech-to-text server on your own hardware. Diction connects to it over your network. Your audio never touches any third-party service. The server ships as a Docker image. One command to start.
iOS requires Full Access for any keyboard extension that uses the network. Diction needs it to send audio to your server or Diction One for transcription. Diction has no QWERTY keys to log, does not read your clipboard, and does not access contacts or any other personal data.
