Aqua Voice: Cross-Application Speech Input to Generate Accurate Text

General Introduction

Aqua Voice is an intelligent speech-based text generation tool focused on quickly converting user speech into formatted text. Founded in 2023 by Finnian Brown and Jack McIntire, Aqua Voice is headquartered in San Francisco, USA, and is part of the Y Combinator W24 incubation program.Aqua Voice not only accurately transcribes speech, but also understands the user's intent based on context and automatically formats text to generate content such as emails, code, or messages. It's responsive, with a startup time of less than 200 milliseconds, a text output latency as low as 450 milliseconds, and an error rate that's about 17 times lower than Siri and Google voice input. With support for Mac and Windows, it can be used in a wide range of applications without additional plug-ins, making it ideal for users who need to get text work done efficiently.

Function List

• High-precision speech transcription: Converts speech to text in real time, automatically correcting spelling, grammar and formatting.
• natural language instruction: Adjust text with simple verbal commands such as "change to list" or "insert table".
• context-sensitive: Intelligently supplement information or optimize output based on screen content or document context.
• ultra-low latency: Start-up time is less than 200 milliseconds and fast mode output delay is about 450 milliseconds.
• streaming mode: Supports continuous voice input with a latency of about 850 milliseconds for complex tasks.
• Cross-application compatibility: Enter text directly into Notion, Slack, VSCode, and other apps without a plugin.
• Code Understanding: Optimize code-related transcription for developers with support for syntax highlighting and terminology correction.
• Custom Dictionary: Add proprietary vocabulary (e.g., personal names, technical terms) to ensure accurate transcription.
• Privacy: Data is processed locally and no user data is stored to safeguard privacy and security.

Using Help

Installation process

1. Visit the official website https://withaqua.com/ and click the "Download" button at the top of the page.
2. Choose the version according to your operating system:
   • Mac users choose the Apple Silicon or Intel version.
   • Windows users download the generic installation package directly.
3. Once the download is complete, double-click on the installation package and follow the prompts to complete the installation. The whole process usually takes only 3-5 minutes.
4. Launch Aqua Voice, the software will automatically detect the microphone and prompt for connection. If you have any problems, you can check the FAQ on the official website for solutions.
5. First-time users need to register for an account, and the free version offers a 1,000-word trial. After the trial, you can choose to subscribe to the Pro version ($10 per month or $96 per year).

How to use

At the core of Aqua Voice is the ability to quickly generate and edit text by voice, which is easy to use and suitable for a wide range of scenarios. Below is a detailed guide to its use:

Basic Voice Input

1. Open Aqua Voice and click on the microphone icon or press the default shortcut key Ctrl+Space(Customizable) Starts recording.
2. Say something like "Write an email to Sarah explaining that tomorrow's meeting is canceled". The software generates the formatted text:

主题：会议取消通知
亲爱的 Sarah，
明天原定的会议已取消，请知悉。谢谢！

3. Short pauses are automatically segmented, and long-pressing the microphone icon ends the recording.

Using Natural Language Instructions

Aqua Voice supports text formatting with simple commands. For example:

• Say "change to list" and the text will change:

- 明天原定的会议已取消
- 请知悉

• Say "Insert Form" to generate it:

| 任务     | 状态   |
|----------|--------|
| 会议     | 取消   |

• When you say "shorten this paragraph", the software will streamline the text, for example, by replacing "Please acknowledge and confirm receipt" with "Please confirm".
  Instructions should be clear and avoid complex statements. For example, "Make this part more concise" is more easily recognized than "Optimize the structure of the text".

Cross-application use

Aqua Voice works in multiple applications without plug-ins:

1. In Slack, Notion, or Gmail, press the shortcut key to activate Aqua Voice.
2. Say something like "Reply to John and tell him the project is complete". The software will enter it directly:

嗨 John，项目已经完成，请确认。

3. Once done, you can send it manually or say "Send" to trigger the in-app send function (app support required).

Featured Function Operation

1. context-sensitive
   Aqua Voice understands context through on-screen content. For example:

• Say "add comment" when writing code, and it will generate something like // 初始化用户数据 The annotations.
• In the email, say "Fill in the date" and it will insert the current date, e.g. "April 10, 2025".
• If a person's name is mentioned (e.g. "Tom"), it will refer to the contact list on the screen to minimize spelling errors.

2. Code Comprehension and Syntax Highlighting
   For developers, Aqua Voice recognizes technical terms and optimizes output:

• Saying "Create function getUserData, accept ID parameter" will generate:

  async function getUserData(id) {
  const response = await fetch(`/users/${id}`);
  return response.json();
  }
  

• Automatically corrects terminology, such as changing "Jason" to "JSON".

3. Streaming mode vs. fast mode

• fast mode(Instant Mode): Suitable for short sentence input with a delay of about 450 milliseconds. The text is output immediately after it is spoken.
• streaming mode(Streaming Mode): for long paragraphs or complex tasks, with a delay of about 850 milliseconds. Generate-as-you-speak, suitable for dictating long documents.
• Switch Mode: Select in the settings, or say "Switch to Streaming Mode".

4. Custom Dictionary

• Add proprietary words such as "Grok" or "xAI" to your settings to ensure accurate transcription.
• Example: After adding "Grok", saying "Grok is an AI assistant" will not be misspelled as "Grock".

5. Privacy and Security

• All voice data is processed locally and not uploaded to the cloud.
• Screen context analysis is only used to optimize the output and no information is stored.

caveat

• Make sure the microphone is of good quality to avoid background noise interfering with transcription accuracy.
• Currently only supports English, Chinese voice input is not supported for the time being, but the development team said it is developing multi-language features.
• Network connectivity improves context-awareness, but offline mode works fine for basic functions.
• Regularly check the official website https://withaqua.com/changelog for the latest updates and the software will automatically prompt for new versions.

advanced skill

• Complex Document Formatting:: Say "Format into report", which generates structured text with headings, body and conclusion.
• multitasking: In streaming mode, say "Write an email to Anna explaining your plans; then create a to-do list" and the software will do it in turn.
• Shortcut Optimization: Adjust the shortcuts in the settings, e.g., by setting the Ctrl+Space change into Alt+V, improving operational efficiency.

With these features, users can easily use their voice to complete edits from simple messages to complex code, dramatically reducing manual input time.

application scenario

1. Effective Communication in the Workplace
   Scenario Description: A busy manager needs to respond to multiple emails in between meetings. Save time by using Aqua Voice to dictate email content and the software automatically generates formatted text that can be sent directly.
2. Rapid coding for developers
   Scenario Description: Programmer dictates code logic, such as "Create REST API endpoint", and Aqua Voice generates the exact code snippet, reducing the need for manual keyboarding.
3. Student classroom notes
   Scenario description: Students record lectures by voice and say "Organize into an outline" to quickly generate review materials for easy organization after class.
4. Accessibility aids
   Scenario Description: Users who cannot type conveniently operate their computers by voice to complete message sending or document editing to enhance their life and work efficiency.

QA

1. Does Aqua Voice support Chinese voice?
   Currently only English is supported, Chinese function is under development. You can follow the official website https://withaqua.com/blog for updates.
2. What are the limitations of the free version?
   The free version offers a 1000 word trial and 5 custom dictionary slots. Unlimited words require a subscription to the Pro version ($10 per month).
3. How to ensure data security?
   Voice and screen data are processed locally, not uploaded to the cloud, and no information is stored without the user's permission.
4. In which applications can it be used?
   Aqua Voice supports Notion, Slack, VSCode, Gmail, WhatsApp, etc. Enter text directly without additional plug-ins.
5. How do you deal with proprietary terms?
   Add custom dictionaries, such as company names or technical terms, to the settings and the software will recognize these words first.