Change from Send button to Enter button on virtual keyboard for Opera AI chat

[Sinraasu]

Sinraasu last edited by

UX Bug: Mobile keyboard shows "Send" instead of "Enter" in Opera AI chat.

When typing a prompt in Opera AI on mobile, the virtual keyboard replaces the standard "Enter/Return" key with a "Send" key (located directly below the backspace button). Because the Send button is so close to Backspace, users frequently and accidentally send incomplete prompts when trying to correct a typo.

Suggested Improvement:

Please update the chat input field's HTML attributes. Changing it to enterkeyhint="enter" or implementing it as a <textarea> will force the mobile keyboard to display a standard "Next Line / Enter" key instead of the "Send" action. Users can then rely on the UI's on-screen send arrow to submit the prompt, drastically reducing accidental submissions.

In the same way, you guys can look into one touch selection option for "Ask More" suggestions. Since the mobile keyboards popup and close down as they please and in different sizes, sometimes I find myself touching one of the Ask More questions even though I didn't want to do it .. but it ends up happening when I try to open the keyboard when I want to start typing a follow up question... so then I have to press the stop button and then type a new question...

P.S. Why this is an easy Energy & Cost Saving Win:

Fixing this keyboard layout isn’t just a UI improvement; it is a direct reduction in unnecessary server load and AI compute costs.

Currently, every time a user accidentally hits "Send" instead of "Backspace":

The incomplete prompt is sent to Opera's servers.

The Large Language Model (LLM) spins up GPU compute power to process the fragment and generate a useless response.

The user has to hit "Stop generating," delete the response, and start typing their prompt all over again.

By simply changing the keyboard key to "Enter" instead of "Send," you instantly eliminate thousands of accidental, useless API calls to the AI model every day. This simple HTML tweak will immediately reduce wasted GPU cycles, lower Opera's API costs, and improve the overall energy efficiency of the Opera AI infrastructure. Pls look into one touch promting of suggested questions as well...

Thank you...!

[tauqirashraf]

tauqirashraf last edited by

This is a good point, especially for users who do most of their work on mobile. Accidentally submitting an incomplete prompt interrupts the workflow and makes the chat experience frustrating.

I also agree about the “Ask More” suggestions. When the keyboard opens or closes, the layout can shift, making it easy to tap a suggestion unintentionally.

One possible solution would be to let users choose whether Enter submits a prompt or inserts a new line, while keeping the send button available as the primary submission control. For suggested questions, a little more spacing or a clearer separation from the input area could help prevent accidental taps.

The potential reduction in unnecessary AI requests is worth investigating too, although the actual cost and energy savings would depend on how frequently these accidental submissions occur.

Hopefully, the Opera team considers both improvements. Small mobile UX changes like these can make a noticeable difference to everyday usability.

[Sinraasu]

Sinraasu @tauqirashraf last edited by

@tauqirashraf thanks for accpeting the idea. I myself have misclicked on the enter or the Ask More options from 10 to 20 times in a day when I'm having a day of intense AI usage... So, even if I press the stop button or ignore the response, I'm sure, Opera will be billed for all those prompt triggers. Even if it is free or for menial cost, I'm sure we can save a lot of energy and unnecessary soft water usage by reducing these kind of prompt wastage... There's a lot of bad PR for ai anyways. We should try to make best and fair use of the technology before it gets to a point where people point fingers or legislate...

[tauqirashraf]

tauqirashraf @Sinraasu last edited by

@Sinraasu I agree with this point. Accidental clicks can trigger unnecessary AI requests, especially for users who rely on these tools throughout the day. Adding a confirmation step or an option to undo a prompt before it starts processing could help reduce wasted requests and resources. Small usability improvements like these can make AI tools more efficient and encourage more responsible usage. Hopefully, the Opera team considers this idea in a future update.