The Impact of AI Voice Input Whisper: Completing Articles Just by Speaking

Ideas you thought you would summarize later end up sinking to the bottom of your memo app.

Have you ever had that experience?

I was like that for a long time. I had ideas, but writing them down was a hassle, so I never ended up sharing them. But it's different now.
Just by speaking, I can create a draft for an article.

Speaking is faster than typing on a keyboard

When using a PC as a hobby.
Writing notes, organizing ideas—that private time is precious, and you want to get through it as efficiently as possible.

I thought I could type on a keyboard as well as anyone. But one day, I realized something.

The speed of my thoughts and the speed of my typing were completely out of sync.

My head is overflowing with words, but my fingers can't keep up. Flick input on a smartphone is even slower. The memos I thought I would clean up later became so fragmented that I didn't even have the energy to read them again.

That's when I tried the AI voice input tool, 'SuperWhisper'.

An era where AI acts as your punctuation clerk

Voice input itself has been around for a long time.

But old voice input had a major flaw. It only transcribed exactly what you said. You had to say the punctuation yourself. Correcting misinterpretations was a hassle, so I ended up going back to the keyboard...

AI voice input is fundamentally different in that regard.

The AI organizes what you say into natural sentences.
It removes hesitations like 'um' and 'uh,' and it adds punctuation based on the context.

In other words, I just have to speak.

While I'm talking, the framework of an article is created.
Talking while on the move turns ideas into memos. That world is already within reach.

What is Whisper?

The 'Whisper' that gives SuperWhisper its name is
a speech recognition AI model developed by OpenAI.

The features of Whisper aremultilingual support and high accuracy.
It supports about 100 languages, including Japanese, and can transcribe relatively accurately even in noisy environments.

The reason I chose a service that uses Whisper is simple: 'Because it is a model specialized and refined for speech recognition among all AI models.' I decided that if I wanted high accuracy, I should use a specialized model.

* Whisper specifications and supported languages are based on information at the time of writing. Please check the official OpenAI website for the latest information.

SuperWhisper vs. Wispr Flow: What's the Difference?

When I researched services using Whisper, two main candidates emerged.

SuperWhisper

• Allows local execution (processing can be completed entirely on your own PC)

• Data does not leak externally, making it suitable for highly confidential use cases

• You can create as many 'personal modes' as you like using custom prompts

• Approx. $8.49/month or a one-time purchase of approx. $250 (*As of writing. Please verify on the official website)

Wispr Flow

• Designed with an emphasis on speed and intuitiveness

• Extremely low latency from speaking to text conversion

• Rich in voice-based editing commands

• Approx. $15/month (approx. $12/month billed annually); free version limited to 2,000 words per week (*As of writing. Please verify on the official website)

I am currently using SuperWhisper.

SuperWhisper has a higher learning curve to master due to its high customizability. However, I believe that the more you can do, the better, so I chose the one that allows for further efficiency improvements.

Creating a 'Personal AI Secretary' with Custom Prompts

This is the true value of SuperWhisper.

Using a feature called Custom Mode, you can finely configure 'how the AI processes what you say' yourself. You can create as many modes as you like and switch between them for different purposes.

For example, you can create modes like these.

📝 note article mode

役割：あなたはnote向けライターです。
ユーザーの話した内容（User Message）を、
読みやすいnote記事の下書きとして整形してください。

要件：
- 語り口は「です・ます調」
- 改行を多用して読みやすく
- 重要な気づきは太字にする
- 「えーと」「あのー」などの言い淀みは除く

This is exactly what I am using right now. What I speak becomes a note draft as is.

📧 email reply mode

ユーザーの話した内容を、丁寧なビジネスメールの文面に変換してください。
件名・本文・署名の構造で出力すること。

Just by saying, 'I want to tell them that next week is fine for that matter,' a proper email draft is generated.

🔄 Japanese-English translation mode

ユーザーの話した日本語を、自然なビジネス英語に翻訳してください。

When I want to leave a note in English, I can just speak in Japanese and it will be output in English.

Even more convenient is the context integration feature, which allows you to pass the 'content of the currently open app' or 'selected text' to the AI for processing.

For example, if you speak while having code selected, instructions like 'add comments to this code' or 'explain this in Japanese' will work immediately.

Just by speaking, the AI automatically handles tasks according to the situation.

This is no longer just 'voice input,' but 'voice instructions to AI.'

The initial hurdle of an English UI

The first thing I stumbled upon when I started using SuperWhisper was that
the settings screen is entirely in English.

Creating modes, entering prompts, setting hotkeys... I had to look things up one by one to understand what changed where.

If you have an allergy to English, you will need a little patience at first. However, the comfort you find after getting through the settings is real.

A world where speaking turns into 'publishing'

What I have realized since I started using AI voice input is that
the hurdle of 'writing' has disappeared.

I have ideas. But writing is a hassle. Because of that gap, there were many things I couldn't publish.

When it became just a matter of speaking, that gap was filled. During the time I spend on my PC for hobbies, just saying what I think becomes output.

Technology is just a tool. But I think this tool is quite effective for people who 'want to publish but are hesitant to start.'

You can think of speaking more lightly.
It was an experience that made me realize that.

If you also feel that writing is a hassle or that you struggle to keep up with your output, please press the like button to let me know. Just knowing that there are others with the same struggle gives me the strength to write the next one.