Have something to say?
Title: Extend or Disable the Voice Input Timeout on Android:
The current voice input duration on Android is too limited for natural speech. Allowing longer listening sessions, or providing an option to adjust the timeout length, would make dictation faster and more efficient.
Enable Automatic Voice Language Detection on Windows and Android
It is recommended to add automatic voice language detection on both the Windows and Android versions of KeyShift AI. This feature would make it easier for users to switch between languages seamlessly, eliminating the need for manual language selection and improving the overall voice input experience.
integration with Prompt Architects
it’s an amazing tool and i’d love to apply it’s very robust features alongside KeyShift. it is standalone but most powerful as an extension
Japanese (IME) support for the typed trigger
Make the "keyshift" typed trigger work with Japanese and other IME-based input.
Preserve embedded images in emails
Rewrite or translate emails with inline images without losing them, keeping the whole text coherent.
Edit before insert
Review and edit the AI result before it goes into the field, including for translations.
A More Visual Way to Switch Between Providers
I’d also love to see a cleaner, more visual interface for switching between providers instead of keeping everything tucked away in a dropdown menu. As the list of available engines grows, a visual selector would make it much easier to compare options at a glance and choose the right one quickly.
More Cloud Engines for Faster, More Accurate Voice Input
I’d love to see more cloud engine options such as Soniox, Groq, and Deepgram. These providers would be especially valuable for users who prioritize extra speed, stronger accuracy, and more reliable transcription in demanding workflows. Having more choices would also let people pick the engine that best matches their needs, whether that’s low latency, multilingual performance, or higher transcription quality for professional use.
On-Device Transcription for Faster, Private Voice Input
On-device transcription should be the default, using local models like Whisper or Parakeet, so the feature remains fast, private, and free to run. This would reduce dependence on cloud processing, improve responsiveness, and give users a more secure experience with their voice data staying on the device.
Inline assistant: Quick Actions near your cursor
Bring the Quick Actions panel right next to where you're working: it opens near your cursor instead of the center of the screen, and after a result you can refine it with follow-up instructions like "shorter" or "more formal", without starting over. Chat-like iteration, always scoped to your selected text.
Save for later: snippets, links & reminders
Select text or copy a link anywhere, and save it to KeyShift with one shortcut. Optionally set a reminder (tonight, tomorrow, or a custom time) and get a native notification that brings the snippet back — with one click to copy it or open the link. Everything is stored locally on your device, in line with our privacy-first approach. Vote if you want this — and tell us in the comments how you'd use it!
Meeting & voice memo summaries
Extending voice dictation: record longer audio — a meeting, a voice memo, a brainstorm — and get a clean transcript with an AI summary and action items.
AI snippets & templates
Reusable text templates with AI-filled variables. Type something like keyshift intro email for [client] and get your own template, personalized by AI, inserted right where you're typing.
Personal writing style
KeyShift learns how you write — your tone, phrasing, and vocabulary — so "make this sound better" produces text that sounds like you, not like a generic AI.
AI Workflows — including your own custom workflows
Chain multiple AI actions into one workflow and run them on your text in a single step — for example: fix grammar → make professional → translate. New: you can now create your own custom workflows and sync them across your devices, not just use the built-in ones.
Voice dictation powered by Whisper
Dictate instead of typing. Your speech is transcribed with Whisper and cleaned up by AI for accurate, natural-sounding text — great for non-native speakers and fast replies.