Have ideas for a shortcut?! Post here!

[cjpais]

This is a thread for collecting ideas for other shortcuts that you might want in handy that right now are not supported, but may be supported in the future when we can figure them out in a nice unified way across the application and code base, as well as in the UI and user experience.

[tachyonicbytes]

The ones from the handy-cli script seemed nice to me. The three remaining ones are:

• generate code
• execute commands
• get AI assistance with context

[adibwoy]

Thank you for building Handy!

Would be great if support for the Right Command key (⌘ right) as transcription trigger could be added

• Avoids conflicts with Left Command shortcuts common in macOS
• Matches VoiceInk's ergonomic approach (macOS distinguishes left/right modifiers)
• Bonus: Allow independent left/right modifier selection in settings

[cjpais]

@adibwoy its a work in progress currently, maybe in 2 weeks it'll be done

[skoch13]

this can be easily done via karabiner elements:

• set right cmd to produce F13
• set F13 as the trigger in Handy

[28Black]

As I suggested here, I would like to see these additional shortcuts:

• Shortcut 1: Standard transcription, as it's now.
• Shortcut 2: Transcription with a space at the end of the sentence, after the period.
• Shortcut 3: Transcription with translation (into English for supported models).

This would allow anyone to simply do what they need to do without interacting with the user interface.
And Handy would become even more handy. 👍

[jkikoo]

Totally agree with these three propositions. +1
I can't wait to have a shortcut to translate to English without having to go into the settings to disable the translation into English toggle.
Thx @cjpais for your work 👍

[anton-averich]

Shortcut 3: Transcription with translation (into English for supported models).
Would be very-very useful!

[schmurtzm]

+1 for Shortcut 3: Transcription with translation (into English for supported models).
Shortcut 2 can easily be done with Replacement feature (#455) (for example create a rule "full stop with space".)

[mvsite]

This sort of a bug report but not for you Mac folks -> on Windows, ctrl+v fails in a lot of situations but shift+insert works universally. For example in ssh terminals.

--- a/src-tauri/src/clipboard.rs
+++ b/src-tauri/src/clipboard.rs
@@
-   #[cfg(target_os = "windows")]
-   let (modifier_key, v_key_code) = (Key::Control, Key::Other(0x56)); // VK_V
+   #[cfg(target_os = "windows")]
+   let (modifier_key, v_key_code) = (Key::Shift, Key::Other(0x2D)); // VK_INSERT

[cjpais]

@mvsite if you want to submit a pr I'm happy to test

[immofon]

It would be useful if users could assign custom keyboard shortcuts. After pressing a key and speaking, Handy would run a user-defined command, provide the spoken text as STDIN, and then insert the text from STDOUT at the cursor position. This would allow users to implement their own logic to process the input.

For example, a user could use this to create a custom voice assistant using an LLM and MCP, leveraging the software's full potential.

[cjpais]

@immofon can you please add the request directly here and close the other discussion. Let's keep it all centralized here

[immofon]

Of course, I have updated the comment and closed the other discussion.

[itissid]

This is a nice proposal. I think from a more high level perspective @immofon can let handy be the orchestrator i.e. how to trigger transcription and leaves the result handling of results to the script.

I am wondering about the reverse scenario i.e. Process controlling Handy. What is handy's philosophy on IPC? e.g. would it prefer a process that wants to control it always send it SIGUSR1/2 as a permanent way to let scripts control handy — or perhaps use named pipes/domain sockets for coordination? Because if we do these then a process can communicate back to handy using IPC for other interactive programs.

[Damocles-fr]

I would love to use mouse keys ! Like Middle+Right click

[cjpais]

mmm yes this is something we should support I think

[wrycode]

self-serve keybinding suggestion

How are you overriding global keybindings across the many actively used Linux environments? If you use XDG Global Shortcuts that should work on recent Wayland versions GNOME and KDE, but there are many others where it will not work.

It may be easier to offer an "outlet valve" in the form of a command that has the same effect as the keybinding. Then all Linux users can bind to that command, no matter what kind of environment they're running.

The command would start/stop transcription - I don't think push-to-talk is viable this way.

[cjpais]

#265 may help some for wayland

[byme8]

Ability to change language depending on the shortcut.
For example, first shortcut will assume that you are speaking in English, second shortcut will assume you speak in other language.

[n4n0r]

Agree !

Hi, thanks for the great work!

I use Handy over multiple languages. Handy get often mixed up when speaking one or multiple languages, sometimes partially translate in English where you should not, or the opposite!

2 different shortcuts for two different languages would make it so much easier to use ! Or even more than two ! (Two is enough for me :) ).

Thanks and keep up the good work !

[jgranduel]

Thanks a lot for Handy ! (discovered this week)
I run it on Windows and default Handy shortcut is in conflict with Windows defined shortcuts. The most annoying one is the selection of Excel Table Columns. Changing it would be useful indeed!

[jgranduel]

Thanks ! Sorry i get it now ! How come I didn't see it before😢!?
The recording is not always triggered, but it might me not mastering the UI yet ! let wait for some more usage.

[cjpais]

what os are you on? if Wayland this is a known issue, please search the existing issues for your issue to see if any apply

[jgranduel]

on Win11 x64.

[ruslan-mogilevskiy]

@cjpais , for me on win11, any shortcut that consists only of 'extension' keys (ctrl+shift, shift+Win) doesnt (trigger Handy) work at all. The 'usual' shortcut, like alt+ works OK.

BTW, thanks for the cool app.

[cjpais]

@ruslan-mogilevskiy yeah only modifiers is not supported currently, but it's the next major refactor/feature I'm working on. Hopefully in a few weeks it will be in. It was taken out of the early builds due to permission issues on macOS

[Puma7]

1. Separate keyboard shortcuts

Add two different hotkeys:

• Shortcut A → Raw transcription
• Shortcut B → Transcription + post-processing

This would allow instant switching based on context without toggling settings each time.

---

2. Multiple post-processing profiles (user-defined)

Let users create several post-processing profiles, for example:

• “Email — professional tone”
• “WhatsApp — casual”
• “Bug report formatter”
• “Summarize”
• “Meeting notes”

---

3. Profile selection before / during / after recording

When starting a post-processed transcription, allow choosing which profile will be used.

Possible UX:

• A small dropdown next to the recording box, showing the currently selected post-processing profile
• Default profile selected automatically
• Optional: Changing the profile while recording updates the one used for the final output

This provides maximum convenience and avoids breaking the flow when switching contexts often.

[jtracey93]

100% supportive of this and came here to post the same thing. I really want an easy way to switch the prompts used in post processing that I have saved, as I have different prompts for different tasks. e.g.

[ruurdboeke]

came here for this as well, this is exactly what I want. I have one for pure improvement, but also one for 'better grammar,

[ArthurReboulSalze]

Shortcut switch on/off for "Translate to English" would be really useful

[NickChristensen]

I would love a smart Push-to-Talk option, where it enables Push-to-Talk mode if the shortcut is held vs pressed. Something like:

• keydown + keyup within n ms: toggle recording until shortcut is pressed again
• keydown, no keyup within n ms: record until keyup

[dannysmith]

I'd like this behaviour too, so long as n ms is small. Other apps that do this cause me to accidentally toggle non PTT-mode all the time because I hit the hotkey, forget what I wanted to say and release it.

I'm happy to look at this along with the whole fn key thing and the rather larger "doing shortcut things in a more extensible, generalised way" thing, once I've ticked off a few other contributions I wanna make.

[ankitdaf]

I would love this feature too, but I think another implementation give more flexibility to users. Maybe we can provide a option for the user to switch between push to talk and continuous mode by adding one more keyboard shortcut. For example, if I have control + space for push to talk, then we could have a control shift space or a control alt space or another such combination for toggling the mode back and forth. I'm happy to pick this up for implementation

[cjpais]

Right now new shortcuts are not going to be merged until we have a nice UI for managing all of the things mentioned in this thread in a unified and not confusing way to people.

[ankitdaf]

@cjpais Are you looking for a design that incorporates all of the configuration parameters in this discussion or is there a list that you're maintaining that you definitely want to include ? I can take a stab at it if you like

[cjpais]

@ankitdaf ideally it would be fairly flexible to let people map the ones they want, but I'm not sure quite yet. Right now I'm mainly working on stabilizing the app and adding new features is generally quite a bit lower on the list. I would love to see proposal UI's rather than code specifically so we can work on that. I don't have much opinion on the specific functionality itself, I mostly just want to see/feel what the UI/UX would be

[g33khubber]

If it's an extra feature, I would love to be able to have a mix of two languages in a sentence. Otherwise, a keyboard shortcut to toggle between two most commonly used languages would be ideal.

[kurisu-dotto-komu]

I'm using Handy on Pop!_OS with COSMIC (Wayland). Since Wayland doesn't allow apps to register global shortcuts the same way X11 does, I'm using the signal approach mentioned in the docs.

For anyone else on COSMIC, you can add a global shortcut by editing:

~/.config/cosmic/com.system76.CosmicSettings.Shortcuts/v1/custom

{
    (
        modifiers: [
            Ctrl,
        ],
        key: "space",
        description: Some("Toggle Handy Recording"),
    ): Spawn("pkill -USR2 -xo handy"),
}

The toggle shortcut works great. However, I'd also like a way to cancel a recording without transcribing/pasting, similar to pressing Escape when the app is focused. Would it be possible to add another signal (e.g. SIGUSR1) for this?

That would allow:

• SIGUSR2 → toggle recording (existing)
• SIGUSR1 → cancel recording (new)

This is what I'd like to do in the future:

(
    modifiers: [
        Ctrl,
        Shift,
    ],
    key: "space",
    description: Some("Cancel Handy Recording"),
): Spawn("pkill -USR1 -xo handy"),

Thank you for the very Handy app.

[jasongitmail]

Please allow a modifier to be the short cut key (without needing to install Karabiner or anything else). In this case, right option key on MacOS.

Example in Paraspeech:

But attempting to set that in Handy gives an error:

[jasongitmail]

One issue I came across: the shortcut is for "option", not "right option" specifically.

Meaning it takes up two keys now and causes a conflict where left option is used as a modifier in some cases (like a modifier to change from zoom in to zoom when viewing an image in Cursor).

It'd be great if it could be specified as specifically right-option or left-option.

[mfyz]

+1 on right-option only way. I use left option for a lot of other stuff, generally navigating words on my input texts. Right option is almost never used which is perfect use case to exclusively assign to handy in my mind.

[cjpais]

Yep it'll probably happen in the next release

[arjunsharma]

+1! I literally never touch my right command or option keys and would love to be able to set one of those to be the trigger key!

[luckman212]

@arjunsharma at least on macOS, as of v0.7.7, this feature is now available

[benedikt-schesch]

What I would really like is a way to make it hit enter after the transcription automatically. The goal is to make any LLM coding tool to start right away after transcribing and no need for me to hit enter

[benhowes]

Came here to post the exact same thing. Being able to press enter would be such a improvement for search and LLM use where post editing your input is generally not required. Who knows you might even want to YOLO on sending DMs without proofing :D

[milkmann]

I use post-processing for this, and in the rules, there is a phrase to go to a new line. This corresponds to pressing Enter and sending a message. It works quite well.

[furan917]

Would it be possible to add the ability to differentiate between left and right cmd/ctrl/shift?

I like to use my right command for transcribe but Handy does not pick up the difference between my left and right like another tool did.

[tachyonicbytes]

This happens for me with handy-keys. Isn't this implemented now? Is it only for option?

[sonyccd]

When changing the keybaord shotcut it should not trigger.

[cjpais]

Agree @sonyccd can you make this an issue, this is a bug

[vruzin]

I want press only one button (left ctrl / right ctrl / left shift / right shift / left alt / right alt) or some media buttons on keyboard.

[aliaksandrsen]

Please consider adding an option to toggle Handy using a double press of a modifier key (for example Left Shift, Right Shift, Ctrl, or Alt).

Behavior:

Double tap -> start speech-to-text
Double tap again -> stop speech-to-text

Benefits:

• Faster and more natural than key combinations
• Reduces shortcut conflicts with other apps
• More convenient during frequent dictation
• Could be optional and configurable in settings

This would provide a simpler and more ergonomic way to control Handy.

For example, the Spokenly app already uses this approach successfully.

[jasonjmcghee]

Strong upvote here - was about to log until I found this.

[Nagell]

I'm not sure if this is a Tauri limitation, but there is no possibility to switch on Windows from default CTRL+SPACE to CTRL+WIN(⊞).
When trying, an error occurs:

[...] Error:
Tauri shortcuts must include a main key (letter,
number, F-key, etc.) in addition to modifiers

CTRL+SPACE is exactly the combination that VS Code is using for IntelliSense by default.
Many other combinations are mostly complicated or already used.
In essence it would be quite practical if possible to just use two "modifiers".

[cjpais]

The Handy Keys keyboard implementation is a work in progress and trying to fix this

[tim-mukhin]

Using F15 on my keyboards (both with and without numpads have them).
On macbook tried to use opt+cmd - not convenient. Just remapped to F5 - it has mic icon on it. Not sure it would work.

But generally it would be great to have the same shortcut on any machine. I really liked using for a couple of days the right control, which I'm not using for anything else. But handy doesn't differ left and right controls. I saw the discussion about adding that and liked it.

[dandzl]

I'd like a keyboard shortcut to "paste last transcript". Right now, I have to move my hand to the mouse, navigate to the menu bar icon, click on copy last transcript, then CMD+V to paste it. this also has the side effect of killing whatever's inside my clipboard already. It'd be great if there was a keyboard shortcut like CMD+Option+V to "paste" what's in the last transcript directly

[cjpais]

I believe you can change the setting to overwrite you clipboard so it's just there, either advanced or debug (cmd+shift+d)

[dandzl]

do you mean change the Clipboard Handling from "Don't Modify Clipboard" to "Copy to Clipboard"? I was hoping we could add an additional shortcut that doesn't overwrite the clipboard, so you can save whatever's on your clipboard, and still paste the last transcript. Hence why I was thinking CMD+Option+V

[juste-un-gars]

Hey! I really love Handy, it's exactly what I was looking for, clean, offline, and works great.

Quick question: is there a way to trigger recording from the command line or an API on Windows? I saw that Linux users can use SIGUSR2 signals, but I couldn't find anything similar for Windows.

I'm trying to use an external Bluetooth button (or other devices) to control recording instead of keyboard shortcuts. The problem is that keyboard shortcuts sometimes fail in certain applications (they get intercepted or blocked), and I'd prefer a more reliable method.

If this doesn't exist yet, would you consider adding it? Something like a simple HTTP endpoint on localhost or a CLI command would be perfect for controlling Handy from external scripts or devices.

Thanks for the awesome work on this project!

[cjpais]

Hey @juste-un-gars, I think we will be adding API support in the nearish future (in the next two months, maybe sooner)

[juste-un-gars]

Superb !

[luckman212]

I'm using Handy 0.7.8 on macOS + Right Command as my shortcut key (not PTT mode)

Right now, the recording starts on keyDown - it should really be on keyUp when not in PTT mode. Because, now I cannot e.g. press ⌘p for "print", ⌘m for "minimize window" etc without also triggering Handy.

[MathieuT1508]

[cjpais]

@MathieuT1508 turn off push to talk in general settings

[MathieuT1508]

It's not working very well; it only detects the phrase I say about half the time. Could you please check it?

If I enable the key press and hold function, then it always works.

I'm using version v0.7.7 = update v0.7.9 , same probleme

[cjpais]

Send a video

[itissid]

Using handy on desktop. Its fantastic! Thanks for making an amazing product. One thing that is noticeable sometimes that after I press the key that records, it can take a a few to several 100 ms for the app/widget to actually start and so the first few words can get missed. It could be that there is latency in the microphone(I use a Insta 360 web cam with speaker) or just a slow machine. This is a UX problem that could be solved in one or more ways e.g.

• play a sound when ready: this is what dictation on mac does(though I am not sure that would worsen latency) because of loading and playing said sound.

• Another idea is for placing a pseudo LED "ready light" in the handy widget and highlight the widget using lighting when the key is pressed — I can focus(just a bit) on the handy widget and see the ready light, especially when I have two monitors active, its hard to know where the handy widget is and if its ready to start transcribing. The motivation for this is something similar to (and a feature I love) in iTerm2 and helps locate the cursor before I start typing, especially in tmux split panes:

Screen.Recording.2026-03-04.at.9.29.55.AM.mov

• Use a feature of accessibility that I don't know about but solves this problem by not coding it in handy.

[itissid]

Another reason to find the handy widget visually might solve (by setting user expectation i.e. not "solving" for complex speech dynamics by implementing this in the handy desktop client code) the use case recording two sentences back to back by pressing shortcut key fairly rapidly, so timing matters:

• Press short cut key for handy -> Handy Widget show -> record voice -> release -> Handy widget shows "transcribing..."
• I immediately press short cut key for recording again, I expecting handy to record -> Widget not active/showing -> I speak -> nothing recorded and transcribed.

[cjpais]

You can enable sound in general which does what you want. The microphone can take time to start.

[juste-un-gars]

I've bought a small usb mic for my Windows machine, with a button on it. It can be detected by windows (sound no sound).
So I was able to create a small app in windows that just detect that to start Handy recording (or stop), and it's working well :).

https://github.com/juste-un-gars/handymic

[itissid]

One suggestion is to have a post processing key to trigger the actions after transcriptions instead of how its done right now via settings, e.g. The Paste Method, Clipboard handling and Auto submit. And it won't be just for AI post processing, but for a number of things e.g. it would be useful to append to the clipboard until I paste it.

A modifier key + one more key changes what you do with the transcribed text. Its sort of like what superkey does with click modes(see screenshot) but for transcribed text. The upshot is having a shortcut for every setting that manipulates teh transcription in some way instead.

The input ergonomics — UI Widget waiting for user input vs disappearing on releasing the shortcut key — might need to be rethought a bit.

I imaging there could be two modes(my transcribe shortcut key is Fn, assume Enter is the post processing key):

• Default(what is now now): Fn Key Press -> record -> Fn key release -> copy to clipboard etc. Text pasting, Clipboard action and submission controlled by the settings.
• Action Mode: Fn Key Press -> record starts(widget is active) -> Fn release transcribes(Widget still active). Enter +0: Does default, Enter+ 1: Engages LLM post processing, Enter +2: appends to clipboard instead of overwriting it, Enter + 3: ...)

[oheidemann]

+1 on this — love using Handy and this touches on my exact pain point.

I use auto-submit as my default and it works great for chat interfaces and LLM tools. But right now there's no quick way to:

1. Cancel a recording before it gets pasted and auto-submitted
2. Skip auto-submit just once to review the transcription before sending

The only option is to go into settings and toggle auto-submit off, which completely breaks the flow.

Your full action mode proposal would definitely cover this and more. But even a simpler solution would go a long way — something like a modifier key when stopping the recording to skip auto-submit, or a second shortcut that transcribes without auto-submitting. No new UI needed, just an alternative path for the cases where you want to review before sending.

[isc]

Paste last transcript shortcut

Use case: I use Handy with push-to-talk while working in a terminal app (Claude Code). While I'm dictating, the terminal sometimes displays a permission prompt that steals focus. When I release push-to-talk, Handy transcribes correctly but can't paste — the active input has changed and the transcription is lost.

My workflow would be:

1. Start dictating (push-to-talk)
2. Terminal prompts for permission mid-dictation → focus shifts
3. Release push-to-talk → Handy transcribes, but paste fails (wrong focus)
4. Handle the permission prompt
5. Press a shortcut to paste the last transcript where I actually need it

This is different from clipboard handling — the transcript may have been transcribed fine, it just couldn't be delivered to the right place. A simple "paste last transcript" shortcut (e.g. Cmd+Shift+V) would solve this cleanly.

This was previously raised in #240 and implemented in #659, but I think the concrete problem wasn't fully captured at the time. It's not about convenience — it's about recovering a transcription that was lost due to a focus change during dictation.

[craftsmandigital]

First of all, I want to say that Handy is an absolutely fantastic application! It works incredibly well, and I use it a lot. I saw this thread and wanted to pitch in with an idea that builds a bit on what others have mentioned above.

Currently, I use a custom post-processing script triggered by a shortcut. It takes my transcribed Norwegian text and translates it into Swedish using a local LLM. While this works, I find the current user interface a bit limiting when it comes to managing and switching between different tasks dynamically.

My idea: Workflow "Profiles" with custom parameters
I would love the UI to support creating specific Workflows or Profiles, where each profile can be assigned to its own custom keyboard shortcut.

Crucially, each shortcut/profile shouldn't just select the Language Model and Post-Processing script, but it should also save specific parameters for both. For example, I might want one shortcut to use a specific LLM model for translation, and another shortcut to use a different LLM model for summarizing.

It is a bit hard to explain with words, so here is an example of how the UI and mapping could conceptually look:

Transcription Model	Model Parameters	Post-Processing step	Post-Processing Parameters	Keyboard Shortcut
Parakeet V2	Default	None	-	Ctrl + i
Parakeet V2	Default	Transcribe rules	rule_level: strict	Ctrl + l + o
Ggml Q5_0 NO Small	lang: no	None	-	Ctrl + i + i
Ggml Q5_0 NO Small	lang: no	Transcribe rules	rule_level: strict	Ctrl + i + o
Ggml Q5_0 NO Small	lang: no	Translate to Swedish	LLM_model: Llama-3, temp: 0.3	Ctrl + i + s

(Note: The shortcuts above are just examples of mapping multiple keys to specific tasks).

I fully understand that if a shortcut triggers a change in the primary language model, it might take some extra time to load the new model into memory before a new transcription can begin. I think users (including myself) are perfectly fine with this slight loading delay, as the massive benefit of having dedicated shortcuts with pre-configured parameters heavily outweighs it.

Thanks for building and maintaining such a great tool!

[remy33]

I second this, but for different reasons.

1. In order for the handy shortcut to work in RDP, I have to assign it to either Ctrl, Alt, Shift or mouse (right ctrl + right alt works well). But it's not really convenient for me. I want also to have another shortcut when I just use my regular PC like 'CTRL + space' so I can still operate conveniently while just using the PC regular.
2. I need a language specific profile, in IT many times I talk with many English terms, setting the language to "Auto" is most of the time results in broken English, where setting language specifically will result in specific language plus the English terms of the subject. For example, it common to say "I need to reach to the IT department to fix my DATA because the REGEX doesn't work. ".
3. Some models are better at peaking English while some are better at other languages, be able to pick a model for language will be very useful (especially when dealing with language specific models).