Voice Coding Software for Developers

More context in. A clearer next step.

Voice coding uses dictation to put spoken instructions into an agent, editor, or terminal. Aqua enters the text; your developer tool generates or changes the code.

[01] Explain
Name the task and its boundaries
Focus the agent’s prompt field. Hold your Aqua key and describe the change, relevant files, constraints, and what a successful result looks like. Release to finish.
[02] Review
Read the prompt before sending
Check file names, paths, technical terms, and negations. Correct the transcript before you submit it. Avoid Send It when you want this review step.
[03] Iterate
Inspect the result, then follow up
Review the diff and run the relevant checks. Speak a focused follow-up with what needs to change. Your normal review and deployment decisions stay in your developer tool.

A useful prompt says more than “build this”

Try the dashboard prompt above. Each part gives the agent information it needs to make a better change.

Describe the behavior
Make a React component called TaskDashboard. Show each task’s title and status, and let me filter by owner.
Add the constraints
Keep the project’s existing design system. Support keyboard navigation. Include an empty state when no tasks match.
Define the check
Test the filter behavior before you finish. After the agent responds, review the diff and confirm that the filter, empty state, and keyboard controls work.

Illustrative prompt shown in the demo above. Aqua transcribes your instructions; the coding agent’s output depends on the tool, repository, and context you provide.

Bring your project’s language with you

Names, libraries, and formatting matter in a prompt. Give Aqua useful context and check the exact text before sending.

Explore the product guides

Aqua’s custom Dictionary for project names and terminology

Add names that are specific to your codebase

Put internal project names, packages, and teammate names in your Dictionary. Keep checking exact paths, identifiers, and flags in the transcript. Configure your Dictionary

Aqua’s writing style and formatting preferences

Set the formatting you prefer

Use Custom Instructions to explain how you want dictated text formatted. Include a short example. The Aqua CLI can also manage these settings from your terminal. See Aqua CLI instructions

Keep the agent handoff explicit

Open your editor or coding agent, place the cursor in its prompt field, and dictate. A natural-language prompt can carry the reasoning behind a change: what the user should experience, what already exists, and what must stay consistent.

When the text arrives, check it before you send. A missing “don’t,” a changed file name, or the wrong number can alter the task. Then inspect the code your agent produces and run the checks that fit the change.

When Deep Context is enabled in desktop Settings, Aqua can use nearby on-screen text to help resolve names and technical vocabulary. Read the Privacy Policy and your organization’s rules before using private material.

Use the keyboard when exact characters matter

Dictate the explanation; type or paste an exact hash, path, flag, or command when one character changes its meaning. Enter credentials through your approved secret-management workflow rather than speaking them aloud.

In a terminal, distinguish an agent’s prompt from a shell prompt. Review the text and arguments before pressing Enter, especially when a command changes files, permissions, packages, or remote systems.

Read the terminal workflow guide.

The evidence behind technical vocabulary

Aqua’s August 2025 AISpeak-10 evaluation measured recognition of English technical terms. The shared chart below shows that published result.

AvalonAqua Voice: 97.4%
Whisper Large v3Open-source reference model: 65.1%

AISpeak-10 is Aqua’s English technical-speech dataset. Avalon transcribed the key term correctly in 97.4% of clips, versus 65.1% for Whisper Large v3. This measures key-term recognition, not overall transcript accuracy or a head-to-head test of these apps. See the evaluation and results.

A separate October 2025 OpenASR report lists 6.24% average word error rate for Avalon v1 English across its benchmark suite. Read the report and methodology. These are historical evaluations, not a promise about every prompt or the current model’s ranking.

Start in an app you already use

Open a text field, place your cursor, and dictate. These guides cover the details for each app.

What Aqua Voice customers say

Trusted by people at

@juanbuis

this is why @aquavoice is better than all other AI transcription apps it actually writes like you talk

@utkusays

just subscribed to aqua voice and wow you guys are cooking, best dictation app i've tried so far

@turboblitzzz

Started using @aquavoice a few weeks ago and now I barely type anything. My housemates want to kick me out because i keep talking to my computer. Life is good.

@ColinHughesUK

Aqua Voice is what Mac dictation should feel like — fast, accurate, natural editing.

@tzmartin

tbh @aquavoice is a vibe coder's delight.. lowest latency, highest quality STT I've used so far

@jamesfong

3 products that have broken into my forever home. And that's not an easy feat for me: @circlebackai, @aquavoice, and @superset_sh

@thebuildguy

Unpaid shout out — @aquavoice is probably one of my most used tool that has saved me quite literally a day of time!

@heyvyner

Mapping @aquavoice to this button on my mouse is a whole different level of productivity.

@nozmen

Aqua Voice: I barely type long prompts now, just talk. Saved ~5 hrs typing in 2 weeks. Much more accurate than macOS dictation.

@curious_vii

Claude Code agents (new CLI view) + Ghostty + @aquavoice is actual calm computing

@gkossakowski

Aqua is a true power tool: quick, reliable and understands your context - they were early at appreciating the value of context engineering.

@ArtemR

Aqua Voice @aquavoice is the voice-based text editor I think all of us have always wanted, but no company has been able to deliver so far. The demo is extremely impressive

@0xferruccio

For coding prompting with voice nothing beats @aquavoice All the other transcription apps don't work well with code like tailwind class names

@jaymehoffman

@aquavoice has saved me over 2 days of typing

@jackson_felty

@aquavoice is a prime example of how much better a product can be when the team focuses on making one simple use case incredible

@rickmanelius

Impressed with @aquavoice accuracy in a noisy airport.

@Memetic_Theory

If you're not using @aquavoice you're moving 15% slower than you could

@chiggyAI

just tried out @aquavoice and by far the most accurate voice to text tool i've used!!

@TheNathanRuff

Been using @aquavoice for a few months. Unbelievably better than the other companies that do this

@avgvstvs96

Aqua Voice is seriously good voice transcription for macOS

Make the workflow your own

Practical guides for the next step.

Go deeper on voice coding

Explore prompting habits and the role of dictation alongside coding agents.

Dictate to Codex CLI

Set up Aqua for prompts in your terminal-based coding workflow.

Work with Claude Code

Give Claude Code spoken instructions and review them before sending.

Use a JetBrains IDE

Bring voice input to your editor’s text fields and AI chat.

Dictate in a terminal

Understand prompt focus, command review, and exact-character inputs.

Configure Aqua from your terminal

Use the Aqua CLI for your Dictionary, instructions, and desktop settings.

Voice coding questions

Does Aqua write code for me?

Aqua turns speech into text. Your coding agent or editor generates or changes code. You review both the prompt and the resulting change.

Which developer tools can I use?

Aqua inserts text in supported text fields across coding agents, editors, and terminals on Mac and Windows. The app guides above cover Cursor, Claude Code, Codex, VS Code, Windsurf, and several terminals. Secure or restricted fields can behave differently.

Can I use Aqua at a terminal prompt?

Yes. Confirm whether you are entering an agent prompt or a shell command. Check the transcript and every argument before running it. Type exact strings when one character changes the result.

How does Aqua handle technical vocabulary?

You can add project names and terms to your Dictionary. With Deep Context enabled, Aqua can use nearby on-screen text to help resolve vocabulary. The dated AISpeak evaluation above gives evidence for technical-term recognition, not a guarantee for every transcript.

Is this the same as using a coding agent’s voice mode?

Aqua is a separate dictation app. It puts spoken text into the field you choose, so you can use the same input workflow across supported editors and terminals. The coding agent still interprets the prompt and performs the work.

Give your coding agent the full thought